This documentation site is for the versions of Synapse maintained by the Matrix.org Foundation (github.com/matrix-org/synapse), available under the Apache 2.0 licence.
Contributions are welcome! Synapse is primarily written in
Python. As a developer, you may be interested in the
following documentation:
Read the Contributing Guide. It is meant
to walk new contributors through the process of developing and submitting a
change to the Synapse codebase (which is hosted on
GitHub).
Want to help keep Synapse going but don't know how to code? Synapse is a
Matrix.org Foundation project. Consider becoming a
supporter on Liberapay,
Patreon or through
PayPal via a one-time donation.
If you are an organisation or enterprise and would like to sponsor development,
reach out to us over email at: support (at) matrix.org
If you've found a security issue in Synapse or any other Matrix.org Foundation
project, please report it to us in accordance with our Security Disclosure
Policy. Thank you!
It is important to choose the name for your server before you install Synapse,
because it cannot be changed later.
The server name determines the "domain" part of user-ids for users on your
server: these will all be of the format @user:my.domain.name. It also
determines how other matrix servers will reach yours for federation.
For a test configuration, set this to the hostname of your server. For a more
production-ready setup, you will probably want to specify your domain
(example.com) rather than a matrix-specific hostname here (in the same way
that your email address is probably user@example.com rather than
user@email.example.com) - but doing so may require more advanced setup: see
Setting up Federation.
Slavi Pantaleev has created an Ansible playbook,
which installs the official Docker image of Matrix Synapse
along with many other Matrix-related services (Postgres database, Element, coturn,
ma1sd, SSL support, etc.).
For more details, see
https://github.com/spantaleev/matrix-docker-ansible-deploy
The fingerprint of the repository signing key (as shown by gpg /usr/share/keyrings/matrix-org-archive-keyring.gpg) is
AAF9AE843A7584B5A3E4CD2BCF45A512DE2DA058.
When installing with Debian packages, you might prefer to place files in
/etc/matrix-synapse/conf.d/ to override your configuration without editing
the main configuration file at /etc/matrix-synapse/homeserver.yaml.
By doing that, you won't be asked if you want to replace your configuration
file when you upgrade the Debian package to a later version.
pip may be outdated (6.0.7-1 and needs to be upgraded to 6.0.8-1 ):
sudo pip install --upgrade pip
If you encounter an error with lib bcrypt causing an Wrong ELF Class:
ELFCLASS32 (x64 Systems), you may need to reinstall py-bcrypt to correctly
compile it under the right architecture. (This should not be needed if
installing under virtualenv):
As of OpenBSD 6.7 Synapse is available as a pre-compiled binary. The filesystem
underlying the homeserver directory (defaults to /var/synapse) has to be
mounted with wxallowed (cf. mount(8)), so creating a separate filesystem
and mounting it to /var/synapse should be taken into consideration.
At least 1GB of free RAM if you want to join large public rooms like #matrix:matrix.org
If building on an uncommon architecture for which pre-built wheels are
unavailable, you will need to have a recent Rust compiler installed. The easiest
way of installing the latest version is to use rustup.
This will download Synapse from PyPI
and install it, along with the python libraries it uses, into a virtual environment
under ~/synapse/env. Feel free to pick a different directory if you
prefer.
This Synapse installation can then be later upgraded by using pip again with the
update flag:
... substituting an appropriate value for --server-name and choosing whether
or not to report usage statistics (hostname, Synapse version, uptime, total
users, etc.) to the developers via the --report-stats argument.
This command will generate you a config file that you can then customise, but it will
also generate a set of keys for you. These keys will allow your homeserver to
identify itself to other homeserver, so don't lose or delete them. It would be
wise to back them up somewhere safe. (If, for whatever reason, you do need to
change your homeserver's keys, you may find that other homeserver have the
old key cached. If you update the signing key, you should change the name of the
key in the <server name>.signing.key file (the second word) to something
different. See the spec for more information on key management).
To actually run your new homeserver, pick a working directory for Synapse to
run (e.g. ~/synapse), and:
Synapse is written in Python but some of the libraries it uses are written in
C. So before we can install Synapse itself we need a working C compiler and the
header files for Python C extensions.
A port of Synapse is available under net/synapse. The filesystem
underlying the homeserver directory (defaults to /var/synapse) has to be
mounted with wxallowed (cf. mount(8)), so creating a separate filesystem
and mounting it to /var/synapse should be taken into consideration.
To be able to build Synapse's dependency on python the WRKOBJDIR
(cf. bsd.port.mk(5)) for building python, too, needs to be on a filesystem
mounted with wxallowed (cf. mount(8)).
Creating a WRKOBJDIR for building python under /usr/local (which on a
default OpenBSD installation is mounted with wxallowed):
doas mkdir /usr/local/pobj_wxallowed
Assuming PORTS_PRIVSEP=Yes (cf. bsd.port.mk(5)) and SUDO=doas are
configured in /etc/mk.conf:
doas chown _pbuild:_pbuild /usr/local/pobj_wxallowed
By default Synapse uses an SQLite database and in doing so trades
performance for convenience. Almost all installations should opt to use PostgreSQL
instead. Advantages include:
significant performance improvements due to the superior threading and
caching model, smarter query optimiser
allowing the DB to be run on separate hardware
For information on how to install and use PostgreSQL in Synapse, please see
Using Postgres
SQLite is only acceptable for testing purposes. SQLite should not be used in
a production server. Synapse will perform poorly when using
SQLite, especially when participating in large rooms.
The default configuration exposes a single HTTP port on the local
interface: http://localhost:8008. It is suitable for local testing,
but for any practical use, you will need Synapse's APIs to be served
over HTTPS.
The recommended way to do so is to set up a reverse proxy on port
8448. You can find documentation on doing so in
the reverse proxy documentation.
Alternatively, you can configure Synapse to expose an HTTPS port. To do
so, you will need to edit homeserver.yaml, as follows:
First, under the listeners option, add the configuration for the
TLS-enabled listener like so:
You will also need to add the options tls_certificate_path and
tls_private_key_path. to your configuration file. You will need to manage provisioning of
these certificates yourself.
You can find more information about these options as well as how to configure synapse in the
configuration manual.
If you are using your own certificate, be sure to use a .pem file that
includes the full certificate chain including any intermediate certificates
(for instance, if using certbot, use fullchain.pem as your certificate, not
cert.pem).
For a more detailed guide to configuring your server for federation, see
Federation.
Setting up the client Well-Known URI is optional but if you set it up, it will
allow users to enter their full username (e.g. @user:<server_name>) into clients
which support well-known lookup to automatically configure the homeserver and
identity server URLs. This is useful so that users don't have to memorize or think
about the actual homeserver URL you are using.
The URL https://<server_name>/.well-known/matrix/client should return JSON in
the following format.
To work in browser based clients, the file must be served with the appropriate
Cross-Origin Resource Sharing (CORS) headers. A recommended value would be
Access-Control-Allow-Origin: * which would allow all browser based clients to
view it.
You should also ensure the public_baseurl option in homeserver.yaml is set
correctly. public_baseurl should be set to the URL that clients will use to
connect to your server. This is the same URL you put for the m.homeserverbase_url above.
It is desirable for Synapse to have the capability to send email. This allows
Synapse to send password reset emails, send verifications when an email address
is added to a user's account, and send email notifications to users when they
receive new messages.
To configure an SMTP server for Synapse, modify the configuration section
headed email, and be sure to have at least the smtp_host, smtp_port
and notif_from fields filled out. You may also need to set smtp_user,
smtp_pass, and require_transport_security.
If email is not configured, password reset, registration and notifications via
email will be disabled.
One way to create a new user is to do so from a client like
Element. This requires registration to be enabled via
the
enable_registration
setting.
Alternatively, you can create new users from the command line. This can be done as follows:
If synapse was installed via pip, activate the virtualenv as follows (if Synapse was
installed via a prebuilt package, register_new_matrix_user should already be
on the search path):
cd ~/synapse
source env/bin/activate
synctl start # if not already running
Run the following command:
register_new_matrix_user -c homeserver.yaml
This will prompt you to add details for the new user, and will then connect to
the running Synapse to create the new user. For example:
New user localpart: erikj
Password:
Confirm password:
Make admin [no]:
Success!
This process uses a setting
registration_shared_secret,
which is shared between Synapse itself and the register_new_matrix_user
script. It doesn't matter what it is (a random value is generated by
--generate-config), but it should be kept secret, as anyone with knowledge of
it can register users, including admin accounts, on your server even if
enable_registration is false.
Synapse includes support for previewing URLs, which is disabled by default. To
turn it on you must enable the url_preview_enabled: True config parameter
and explicitly specify the IP ranges that Synapse is not allowed to spider for
previewing in the url_preview_ip_range_blacklist configuration parameter.
This is critical from a security perspective to stop arbitrary Matrix users
spidering 'internal' URLs on your network. At the very least we recommend that
your loopback and RFC1918 IP addresses are blacklisted.
This also requires the optional lxml python dependency to be installed. This
in turn requires the libxml2 library to be available - on Debian/Ubuntu this
means apt-get install libxml2-dev, or equivalent for your OS.
pip seems to leak lots of memory during installation. For instance, a Linux
host with 512MB of RAM may run out of memory whilst installing Twisted. If this
happens, you will have to individually install the dependencies which are
failing, e.g.:
Synapse will require the python postgres client library in order to
connect to a postgres database.
If you are using the matrix.org debian/ubuntu
packages, the necessary python
library will already be installed, but you will need to ensure the
low-level postgres library is installed, which you can do with
apt install libpq5.
For other pre-built packages, please consult the documentation from
the relevant package.
If you installed synapse in a
virtualenv, you can install
the library with:
(substituting the path to your virtualenv for ~/synapse/env, if
you used a different path). You will require the postgres
development files. These are in the libpq-dev package on
Debian-derived distributions.
Assuming your PostgreSQL database user is called postgres, first authenticate as the database user with:
su - postgres
# Or, if your system uses sudo to get administrative rights
sudo -u postgres bash
Then, create a postgres user and a database with:
# this will prompt for a password for the new user
createuser --pwprompt synapse_user
createdb --encoding=UTF8 --locale=C --template=template0 --owner=synapse_user synapse
The above will create a user called synapse_user, and a database called
synapse.
Note that the PostgreSQL database must have the correct encoding set
(as shown above), otherwise it will not be able to store UTF8 strings.
All key, values in args are passed to the psycopg2.connect(..)
function, except keys beginning with cp_, which are consumed by the
twisted adbapi connection pool. See the libpq
documentation
for a list of options which can be passed.
You should consider tuning the args.keepalives_* options if there is any danger of
the connection between your homeserver and database dropping, otherwise Synapse
may block for an extended period while it waits for a response from the
database server. Example values might be:
database:
args:
# ... as above
# seconds of inactivity after which TCP should send a keepalive message to the server
keepalives_idle: 10
# the number of seconds after which a TCP keepalive message that is not
# acknowledged by the server should be retransmitted
keepalives_interval: 10
# the number of TCP keepalives that can be lost before the client's connection
# to the server is considered dead
keepalives_count: 3
In particular, we've found tuning the following values helpful for
performance:
shared_buffers
effective_cache_size
work_mem
maintenance_work_mem
autovacuum_work_mem
Note that the appropriate values for those fields depend on the amount
of free memory the database host has available.
Additionally, admins of large deployments might want to consider using huge pages
to help manage memory, especially when using large values of shared_buffers. You
can read more about that here.
The script synapse_port_db allows porting an existing synapse server
backed by SQLite to using PostgreSQL. This is done in as a two phase
process:
Copy the existing SQLite database to a separate location and run
the port script against that offline database.
Shut down the server. Rerun the port script to port any data that
has come in since taking the first snapshot. Restart server against
the PostgreSQL database.
The port script is designed to be run repeatedly against newer snapshots
of the SQLite database file. This makes it safe to repeat step 1 if
there was a delay between taking the previous snapshot and being ready
to do step 2.
It is safe to at any time kill the port script and restart it.
However, under no circumstances should the SQLite database be VACUUMed between
multiple runs of the script. Doing so can lead to an inconsistent copy of your database
into Postgres.
To avoid accidental error, the script will check that SQLite's auto_vacuum mechanism
is disabled, but the script is not able to protect against a manual VACUUM operation
performed either by the administrator or by any automated task that the administrator
may have configured.
Note that the database may take up significantly more (25% - 100% more)
space on disk after porting to Postgres.
Firstly, shut down the currently running synapse server and copy its
database file (typically homeserver.db) to another location. Once the
copy is complete, restart synapse. For instance:
The flag --curses displays a coloured curses progress UI.
If the script took a long time to complete, or time has otherwise passed
since the original snapshot was taken, repeat the previous steps with a
newer snapshot.
To complete the conversion shut down the synapse server and run the port
script one last time, e.g. if the SQLite database is at homeserver.db
run:
If you get an error along the lines of FATAL: Ident authentication failed for user "synapse_user", you may need to use an authentication method other than
ident:
If the synapse_user user has a password, add the password to the database:
section of homeserver.yaml. Then add the following to pg_hba.conf:
host synapse synapse_user ::1/128 md5 # or `scram-sha-256` instead of `md5` if you use that
If the synapse_user user does not have a password, then a password doesn't
have to be added to homeserver.yaml. But the following does need to be added
to pg_hba.conf:
host synapse synapse_user ::1/128 trust
Note that line order matters in pg_hba.conf, so make sure that if you do add a
new line, it is inserted before:
Synapse will refuse to set up a new database if it has the wrong values of
COLLATE and CTYPE set. Synapse will also refuse to start an existing database with incorrect values
of COLLATE and CTYPE unless the config flag allow_unsafe_locale, found in the
database section of the config, is set to true. Using different locales can cause issues if the locale library is updated from
underneath the database, or if a different version of the locale is used on any
replicas.
If you have a database with an unsafe locale, the safest way to fix the issue is to dump the database and recreate it with
the correct locale parameter (as shown above). It is also possible to change the
parameters on a live database and run a REINDEX on the entire database,
however extreme care must be taken to avoid database corruption.
Note that the above may fail with an error about duplicate rows if corruption
has already occurred, and such duplicate rows will need to be manually removed.
Synapse uses Postgres sequences to generate IDs for various tables. A sequence
and associated table can get out of sync if, for example, Synapse has been
downgraded and then upgraded again.
To fix the issue shut down Synapse (including any and all workers) and run the
SQL command included in the error message. Once done Synapse should start
successfully.
It is recommended to put a reverse proxy such as
nginx,
Apache,
Caddy,
HAProxy or
relayd in front of Synapse. One advantage
of doing so is that it means that you can expose the default https port
(443) to Matrix clients without needing to run Synapse with root
privileges.
You should configure your reverse proxy to forward requests to /_matrix or
/_synapse/client to Synapse, and have it set the X-Forwarded-For and
X-Forwarded-Proto request headers.
You should remember that Matrix clients and other Matrix servers do not
necessarily need to connect to your server via the same server name or
port. Indeed, clients will use port 443 by default, whereas servers default to
port 8448. Where these are different, we refer to the 'client port' and the
'federation port'. See the Matrix
specification
for more details of the algorithm used for federation connections, and
Delegation for instructions on setting up delegation.
NOTE: Your reverse proxy must not canonicalise or normalise
the requested URI in any way (for example, by decoding %xx escapes).
Beware that Apache will canonicalise URIs unless you specify
nocanon.
Let's assume that we expect clients to connect to our server at
https://matrix.example.com, and other servers to connect at
https://example.com:8448. The following sections detail the configuration of
the reverse proxy and the homeserver.
The HTTP configuration will need to be updated for Synapse to correctly record
client IP addresses and generate redirect URLs while behind a reverse proxy.
In homeserver.yaml set x_forwarded: true in the port 8008 section and
consider setting bind_addresses: ['127.0.0.1'] so that the server only
listens to traffic on localhost. (Do not change bind_addresses to 127.0.0.1
when using a containerized Synapse, as that will prevent it from responding
to proxied traffic.)
Optionally, you can also set
request_id_header
so that the server extracts and re-uses the same request ID format that the
reverse proxy is using.
server {
listen 443 ssl http2;
listen [::]:443 ssl http2;
# For the federation port
listen 8448 ssl http2 default_server;
listen [::]:8448 ssl http2 default_server;
server_name matrix.example.com;
location ~ ^(/_matrix|/_synapse/client) {
# note: do not add a path (even a single /) after the port in `proxy_pass`,
# otherwise nginx will canonicalise the URI and cause signature verification
# errors.
proxy_pass http://localhost:8008;
proxy_set_header X-Forwarded-For $remote_addr;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header Host $host;
# Nginx by default only allows file uploads up to 1M in size
# Increase client_max_body_size to match max_upload_size defined in homeserver.yaml
client_max_body_size 50M;
# Synapse responses may be chunked, which is an HTTP/1.1 feature.
proxy_http_version 1.1;
}
}
<VirtualHost *:443>
SSLEngine on
ServerName matrix.example.com
RequestHeader set "X-Forwarded-Proto" expr=%{REQUEST_SCHEME}
AllowEncodedSlashes NoDecode
ProxyPreserveHost on
ProxyPass /_matrix http://127.0.0.1:8008/_matrix nocanon
ProxyPassReverse /_matrix http://127.0.0.1:8008/_matrix
ProxyPass /_synapse/client http://127.0.0.1:8008/_synapse/client nocanon
ProxyPassReverse /_synapse/client http://127.0.0.1:8008/_synapse/client
</VirtualHost>
<VirtualHost *:8448>
SSLEngine on
ServerName example.com
RequestHeader set "X-Forwarded-Proto" expr=%{REQUEST_SCHEME}
AllowEncodedSlashes NoDecode
ProxyPass /_matrix http://127.0.0.1:8008/_matrix nocanon
ProxyPassReverse /_matrix http://127.0.0.1:8008/_matrix
</VirtualHost>
NOTE: ensure the nocanon options are included.
NOTE 2: It appears that Synapse is currently incompatible with the ModSecurity module for Apache (mod_security2). If you need it enabled for other services on your web server, you can disable it for Synapse's two VirtualHosts by including the following lines before each of the two </VirtualHost> above:
<IfModule security2_module>
SecRuleEngine off
</IfModule>
NOTE 3: Missing ProxyPreserveHost on can lead to a redirect loop.
table <webserver> { 127.0.0.1 }
table <matrixserver> { 127.0.0.1 }
http protocol "https" {
tls { no tlsv1.0, ciphers "HIGH" }
tls keypair "example.com"
match header set "X-Forwarded-For" value "$REMOTE_ADDR"
match header set "X-Forwarded-Proto" value "https"
# set CORS header for .well-known/matrix/server, .well-known/matrix/client
# httpd does not support setting headers, so do it here
match request path "/.well-known/matrix/*" tag "matrix-cors"
match response tagged "matrix-cors" header set "Access-Control-Allow-Origin" value "*"
pass quick path "/_matrix/*" forward to <matrixserver>
pass quick path "/_synapse/client/*" forward to <matrixserver>
# pass on non-matrix traffic to webserver
pass forward to <webserver>
}
relay "https_traffic" {
listen on egress port 443 tls
protocol "https"
forward to <matrixserver> port 8008 check tcp
forward to <webserver> port 8080 check tcp
}
http protocol "matrix" {
tls { no tlsv1.0, ciphers "HIGH" }
tls keypair "example.com"
block
pass quick path "/_matrix/*" forward to <matrixserver>
pass quick path "/_synapse/client/*" forward to <matrixserver>
}
relay "matrix_federation" {
listen on egress port 8448 tls
protocol "matrix"
forward to <matrixserver> port 8008 check tcp
}
Synapse exposes a health check endpoint for use by reverse proxies.
Each configured HTTP listener has a /health endpoint which always returns
200 OK (and doesn't get logged).
Endpoints for administering your Synapse instance are placed under
/_synapse/admin. These require authentication through an access token of an
admin user. However as access to these endpoints grants the caller a lot of power,
we do not recommend exposing them to the public internet without good reason.
You can use Synapse with a forward or outbound proxy. An example of when
this is necessary is in corporate environments behind a DMZ (demilitarized zone).
Synapse supports routing outbound HTTP(S) requests via a proxy. Only HTTP(S)
proxy is supported, not SOCKS proxy or anything else.
The http_proxy, https_proxy, no_proxy environment variables are used to
specify proxy settings. The environment variable is not case sensitive.
http_proxy: Proxy server to use for HTTP requests.
https_proxy: Proxy server to use for HTTPS requests.
no_proxy: Comma-separated list of hosts, IP addresses, or IP ranges in CIDR
format which should not use the proxy. Synapse will directly connect to these hosts.
The http_proxy and https_proxy environment variables have the form: [scheme://][<username>:<password>@]<host>[:<port>]
Supported schemes are http:// and https://. The default scheme is http://
for compatibility reasons; it is recommended to set a scheme. If scheme is set
to https:// the connection uses TLS between Synapse and the proxy.
NOTE: Synapse validates the certificates. If the certificate is not
valid, then the connection is dropped.
Default port if not given is 1080.
Username and password are optional and will be used to authenticate against
the proxy.
NOTE:
Synapse does not apply the IP blacklist to connections through the proxy (since
the DNS resolution is done by the proxy). It is expected that the proxy or firewall
will apply blacklisting of IP addresses.
If a proxy server is used with TLS (HTTPS) and no connections are established,
it is most likely due to the proxy's certificates. To test this, the validation
in Synapse can be deactivated.
NOTE: This has an impact on security and is for testing purposes only!
To deactivate the certificate validation, the following setting must be added to
your homserver.yaml.
This document explains how to enable VoIP relaying on your homeserver with
TURN.
The synapse Matrix homeserver supports integration with TURN server via the
TURN server REST API. This
allows the homeserver to generate credentials that are valid for use on the
TURN server through the use of a secret shared between the homeserver and the
TURN server.
This documentation provides two TURN server configuration examples:
For TURN relaying to work, the TURN service must be hosted on a server/endpoint with a public IP.
Hosting TURN behind NAT requires port forwarding and for the NAT gateway to have a public IP.
However, even with appropriate configuration, NAT is known to cause issues and to often not work.
Afterwards, the homeserver needs some further configuration.
As an example, here is the relevant section of the config file for matrix.org. The
turn_uris are appropriate for TURN servers listening on the default ports, with no TLS.
The normal symptoms of a misconfigured TURN server are that calls between
devices on different networks ring, but get stuck at "call
connecting". Unfortunately, troubleshooting this can be tricky.
Here are a few things to try:
Check that you have opened your firewall to allow TCP and UDP traffic to the
TURN ports (normally 3478 and 5349).
Check that you have opened your firewall to allow UDP traffic to the UDP
relay ports (49152-65535 by default).
Try disabling TLS/DTLS listeners and enable only its (unencrypted)
TCP/UDP listeners. (This will only leave signaling traffic unencrypted;
voice & video WebRTC traffic is always encrypted.)
Some WebRTC implementations (notably, that of Google Chrome) appear to get
confused by TURN servers which are reachable over IPv6 (this appears to be
an unexpected side-effect of its handling of multiple IP addresses as
defined by
draft-ietf-rtcweb-ip-handling).
Try removing any AAAA records for your TURN server, so that it is only
reachable over IPv4.
If your TURN server is behind NAT:
double-check that your NAT gateway is correctly forwarding all TURN
ports (normally 3478 & 5349 for TCP & UDP TURN traffic, and 49152-65535 for the UDP
relay) to the NAT-internal address of your TURN server. If advertising
both IPv4 and IPv6 external addresses via the external-ip option, ensure
that the NAT is forwarding both IPv4 and IPv6 traffic to the IPv4 and IPv6
internal addresses of your TURN server. When in doubt, remove AAAA records
for your TURN server and specify only an IPv4 address as your external-ip.
ensure that your TURN server uses the NAT gateway as its default route.
Enable more verbose logging, in coturn via the verbose setting:
verbose
or with eturnal with the shell command eturnalctl loglevel debug or in the configuration file (the service needs to reload for it to become effective):
## Logging configuration:
log_level: debug
... and then see if there are any clues in its logs.
If you are using a browser-based client under Chrome, check
chrome://webrtc-internals/ for insights into the internals of the
negotiation. On Firefox, check the "Connection Log" on about:webrtc.
(Understanding the output is beyond the scope of this document!)
You can test your Matrix homeserver TURN setup with https://test.voip.librepush.net/.
Note that this test is not fully reliable yet, so don't be discouraged if
the test fails.
Here is the github repo of the
source of the tester, where you can file bug reports.
look for the GET /_matrix/client/r0/voip/turnServer request made by a
matrix client to your homeserver in your browser's network inspector. In
the response you should see username and password. Or:
Or (coturn only): Temporarily configure coturn to accept a static
username/password. To do this, comment out use-auth-secret and
static-auth-secret and add the following:
lt-cred-mech
user=username:password
Note: these settings will not take effect unless use-auth-secret
and static-auth-secret are disabled.
Restart coturn after changing the configuration file.
Remember to restore the original settings to go back to testing with
Matrix clients!
If the TURN server is working correctly, you should see at least one relay
entry in the results.
Download the latest release from github. Unpack it and cd into the directory.
Configure it:
./configure
You may need to install libevent2: if so, you should do so in
the way recommended by your operating system. You can ignore
warnings about lack of database support: a database is unnecessary
for this purpose.
See turnserver.conf for explanations of the options. One way to generate
the static-auth-secret is with pwgen:
pwgen -s 64 1
A realm must be specified, but its value is somewhat arbitrary. (It is
sent to clients as part of the authentication flow.) It is conventional to
set it to be your server name.
You will most likely want to configure coturn to write logs somewhere. The
easiest way is normally to send them to the syslog:
syslog
(in which case, the logs will be available via journalctl -u coturn on a
systemd system). Alternatively, coturn can be configured to write to a
logfile - check the example config file supplied with coturn.
Consider your security settings. TURN lets users request a relay which will
connect to arbitrary IP addresses and ports. The following configuration is
suggested as a minimum starting point:
# VoIP traffic is all UDP. There is no reason to let users connect to arbitrary TCP endpoints via the relay.
no-tcp-relay
# don't let the relay ever try to connect to private IP address ranges within your network (if any)
# given the turn server is likely behind your firewall, remember to include any privileged public IPs too.
denied-peer-ip=10.0.0.0-10.255.255.255
denied-peer-ip=192.168.0.0-192.168.255.255
denied-peer-ip=172.16.0.0-172.31.255.255
# recommended additional local peers to block, to mitigate external access to internal services.
# https://www.rtcsec.com/article/slack-webrtc-turn-compromise-and-bug-bounty/#how-to-fix-an-open-turn-relay-to-address-this-vulnerability
no-multicast-peers
denied-peer-ip=0.0.0.0-0.255.255.255
denied-peer-ip=100.64.0.0-100.127.255.255
denied-peer-ip=127.0.0.0-127.255.255.255
denied-peer-ip=169.254.0.0-169.254.255.255
denied-peer-ip=192.0.0.0-192.0.0.255
denied-peer-ip=192.0.2.0-192.0.2.255
denied-peer-ip=192.88.99.0-192.88.99.255
denied-peer-ip=198.18.0.0-198.19.255.255
denied-peer-ip=198.51.100.0-198.51.100.255
denied-peer-ip=203.0.113.0-203.0.113.255
denied-peer-ip=240.0.0.0-255.255.255.255
# special case the turn server itself so that client->TURN->TURN->client flows work
# this should be one of the turn server's listening IPs
allowed-peer-ip=10.0.0.1
# consider whether you want to limit the quota of relayed streams per user (or total) to avoid risk of DoS.
user-quota=12 # 4 streams per video call, so 12 streams = 3 simultaneous relayed calls per user.
total-quota=1200
Also consider supporting TLS/DTLS. To do this, add the following settings
to turnserver.conf:
# TLS certificates, including intermediate certs.
# For Let's Encrypt certificates, use `fullchain.pem` here.
cert=/path/to/fullchain.pem
# TLS private key file
pkey=/path/to/privkey.pem
# Ensure the configuration lines that disable TLS/DTLS are commented-out or removed
#no-tls
#no-dtls
In this case, replace the turn: schemes in the turn_uris settings below
with turns:.
We recommend that you only try to set up TLS/DTLS once you have set up a
basic installation and got it working.
NB: If your TLS certificate was provided by Let's Encrypt, TLS/DTLS will
not work with any Matrix client that uses Chromium's WebRTC library. This
currently includes Element Android & iOS; for more details, see their
respectiveissues as well as the underlying
WebRTC issue.
Consider using a ZeroSSL certificate for your TURN server as a working alternative.
Ensure your firewall allows traffic into the TURN server on the ports
you've configured it to listen on (By default: 3478 and 5349 for TURN
traffic (remember to allow both TCP and UDP traffic), and ports 49152-65535
for the UDP relay.)
If your TURN server is behind NAT, the NAT gateway must have an external,
publicly-reachable IP address. You must configure coturn to advertise that
address to connecting clients:
external-ip=EXTERNAL_NAT_IPv4_ADDRESS
You may optionally limit the TURN server to listen only on the local
address that is mapped by NAT to the external address:
listening-ip=INTERNAL_TURNSERVER_IPv4_ADDRESS
If your NAT gateway is reachable over both IPv4 and IPv6, you may
configure coturn to advertise each available address:
When advertising an external IPv6 address, ensure that the firewall and
network settings of the system running your TURN server are configured to
accept IPv6 traffic, and that the TURN server is listening on the local
IPv6 address that is mapped by NAT to the external IPv6 address.
(Re)start the turn server:
If you used the Debian package (or have set up a systemd unit yourself):
The eturnal TURN server implementation is available from a variety of sources
such as native package managers, binary packages, installation from source or
container image. They are
all described here.
Quick-Test instructions in a Linux Shell
or with Docker
are available as well.
After installation, eturnal usually ships a default configuration file
here: /etc/eturnal.yml (and, if not found there, there is a backup file here:
/opt/eturnal/etc/eturnal.yml). It uses the (indentation-sensitive!) YAML
format. The file contains further explanations.
eturnal runs out of the box with the default configuration. To enable TURN and
to integrate it with your homeserver, some aspects in eturnal's default configuration file
must be edited:
Homeserver's turn_shared_secret
and eturnal's shared secret for authentication
Both need to have the same value. Uncomment and adjust this line in eturnal's
configuration file:
If your TURN server is behind NAT, the NAT gateway must have an external,
publicly-reachable IP address. eturnal tries to autodetect the public IP address,
however, it may also be configured by uncommenting and adjusting this line, so
eturnal advertises that address to connecting clients:
relay_ipv4_addr: "203.0.113.4" # The server's public IPv4 address.
If your NAT gateway is reachable over both IPv4 and IPv6, you may
configure eturnal to advertise each available address:
relay_ipv4_addr: "203.0.113.4" # The server's public IPv4 address.
relay_ipv6_addr: "2001:db8::4" # The server's public IPv6 address (optional).
When advertising an external IPv6 address, ensure that the firewall and
network settings of the system running your TURN server are configured to
accept IPv6 traffic, and that the TURN server is listening on the local
IPv6 address that is mapped by NAT to the external IPv6 address.
Logging
If eturnal was started by systemd, log files are written into the
/var/log/eturnal directory by default. In order to log to the journal
instead, the log_dir option can be set to stdout in the configuration file.
Security considerations
Consider your security settings. TURN lets users request a relay which will
connect to arbitrary IP addresses and ports. The following configuration is
suggested as a minimum starting point, see also the official documentation:
## Reject TURN relaying from/to the following addresses/networks:
blacklist: # This is the default blacklist.
- "127.0.0.0/8" # IPv4 loopback.
- "::1" # IPv6 loopback.
- recommended # Expands to a number of networks recommended to be
# blocked, but includes private networks. Those
# would have to be 'whitelist'ed if eturnal serves
# local clients/peers within such networks.
To whitelist IP addresses or specific (private) networks, you need to add a
whitelist part into the configuration file, e.g.:
Also consider supporting TLS/DTLS. To do this, adjust the following settings
in the eturnal.yml configuration file (TLS parts should not be commented anymore):
In this case, replace the turn: schemes in homeserver's turn_uris settings
with turns:. More is described here.
We recommend that you only try to set up TLS/DTLS once you have set up a
basic installation and got it working.
NB: If your TLS certificate was provided by Let's Encrypt, TLS/DTLS will
not work with any Matrix client that uses Chromium's WebRTC library. This
currently includes Element Android & iOS; for more details, see their
respectiveissues as well as the underlying
WebRTC issue.
Consider using a ZeroSSL certificate for your TURN server as a working alternative.
Firewall
Ensure your firewall allows traffic into the TURN server on the ports
you've configured it to listen on (By default: 3478 and 5349 for TURN
traffic (remember to allow both TCP and UDP traffic), and ports 49152-65535
for the UDP relay.)
Reload/ restarting eturnal
Changes in the configuration file require eturnal to reload/ restart, this
can be achieved by:
eturnalctl reload
eturnal performs a configuration check before actually reloading/ restarting
and provides hints, if something is not correctly configured.
eturnal offers a handy operations script
which can be called e.g. to check, whether the service is up, to restart the service,
to query how many active sessions exist, to change logging behaviour and so on.
Hint: If eturnalctl is not part of your $PATH, consider either sym-linking it (e.g. ´ln -s /opt/eturnal/bin/eturnalctl /usr/local/bin/eturnalctl´) or call it from the default eturnal directory directly: e.g. /opt/eturnal/bin/eturnalctl info
In the following documentation, we use the term server_name to refer to that setting
in your homeserver configuration file. It appears at the ends of user ids, and tells
other homeservers where they can find your server.
By default, other homeservers will expect to be able to reach yours via
your server_name, on port 8448. For example, if you set your server_name
to example.com (so that your user names look like @user:example.com),
other servers will try to connect to yours at https://example.com:8448/.
Delegation is a Matrix feature allowing a homeserver admin to retain a
server_name of example.com so that user IDs, room aliases, etc continue
to look like *:example.com, whilst having federation traffic routed
to a different server and/or port (e.g. synapse.example.com:443).
To use this method, you need to be able to configure the server at
https://<server_name> to serve a file at
https://<server_name>/.well-known/matrix/server. There are two ways to do this, shown below.
Note that the .well-known file is hosted on the default port for https (port 443).
For maximum flexibility, you need to configure an external server such as nginx, Apache
or HAProxy to serve the https://<server_name>/.well-known/matrix/server file. Setting
up such a server is out of the scope of this documentation, but note that it is often
possible to configure your reverse proxy for this.
The URL https://<server_name>/.well-known/matrix/server should be configured
return a JSON structure containing the key m.server like this:
In our example (where we want federation traffic to be routed to
https://synapse.example.com, on port 443), this would mean that
https://example.com/.well-known/matrix/server should return:
{
"m.server": "synapse.example.com:443"
}
Note, specifying a port is optional. If no port is specified, then it defaults
to 8448.
If you are able to set up your domain so that https://<server_name> is routed to
Synapse (i.e., the only change needed is to direct federation traffic to port 443
instead of port 8448), then it is possible to configure Synapse to serve a suitable
.well-known/matrix/server file. To do so, add the following to your homeserver.yaml
file:
serve_server_wellknown: true
Note: this only works if https://<server_name> is routed to Synapse, so is
generally not suitable if Synapse is hosted at a subdomain such as
https://synapse.example.com.
It is also possible to do delegation using a SRV DNS record. However, that is generally
not recommended, as it can be difficult to configure the TLS certificates correctly in
this case, and it offers little advantage over .well-known delegation.
Please keep in mind that server delegation is a function of server-server communication,
and as such using SRV DNS records will not cover use cases involving client-server comms.
This means setting global client settings (such as a Jitsi endpoint, or disabling
creating new rooms as encrypted by default, etc) will still require that you serve a file
from the https://<server_name>/.well-known/ endpoints defined in the spec! If you are
considering using SRV DNS delegation to avoid serving files from this endpoint, consider
the impact that you will not be able to change those client-based default values globally,
and will be relegated to the featureset of the configuration of each individual client.
However, if you really need it, you can find some documentation on what such a
record should look like and how Synapse will use it in the Matrix
specification.
If your homeserver's APIs are accessible on the default federation port (8448)
and the domain your server_name points to, you do not need any delegation.
For instance, if you registered example.com and pointed its DNS A record at a
fresh server, you could install Synapse on that host, giving it a server_name
of example.com, and once a reverse proxy has been set up to proxy all requests
sent to the port 8448 and serve TLS certificates for example.com, you
wouldn't need any delegation set up.
However, if your homeserver's APIs aren't accessible on port 8448 and on the
domain server_name points to, you will need to let other servers know how to
find it using delegation.
Generally, using a reverse proxy for both the federation and client traffic is a good
idea, since it saves handling TLS traffic in Synapse. See
the reverse proxy documentation for information on setting up a
reverse proxy.
Before upgrading check if any special steps are required to upgrade from
the version you currently have installed to the current version of
Synapse. The extra instructions that may be required are listed later in
this document.
Check that your versions of Python and PostgreSQL are still
supported.
Synapse follows upstream lifecycles for Python and
PostgreSQL, and removes support for versions
which are no longer maintained.
If Synapse was installed using prebuilt packages,
you will need to follow the normal process for upgrading those packages.
If Synapse was installed using pip then upgrade to the latest
version by running:
pip install --upgrade matrix-synapse
If Synapse was installed from source, then:
Obtain the latest version of the source code. Git users can run
git pull to do this.
If you're running Synapse in a virtualenv, make sure to activate it before
upgrading. For example, if Synapse is installed in a virtualenv in ~/synapse/env then
run:
Include any relevant extras between square brackets, e.g. pip install --upgrade ".[postgres,oidc]".
If you're using poetry to manage a Synapse installation, run:
poetry install
Include any relevant extras with --extras, e.g. poetry install --extras postgres --extras oidc.
It's probably easiest to run poetry install --extras all.
Restart Synapse:
synctl restart
To check whether your update was successful, you can check the running
server version with:
# you may need to replace 'localhost:8008' if synapse is not configured
# to listen on port 8008.
curl http://localhost:8008/_synapse/admin/v1/server_version
Rolling back to previous releases can be difficult, due to database
schema changes between releases. Where we have been able to test the
rollback process, this will be noted below.
In general, you will need to undo any changes made during the upgrade
process, for example:
Generally Synapse database schemas are compatible across multiple versions, once
a version of Synapse is deployed you may not be able to rollback automatically.
The following table gives the version ranges and the earliest version they can
be rolled back to. E.g. Synapse versions v1.58.0 through v1.61.1 can be rolled
back safely to v1.57.0, but starting with v1.62.0 it is only safe to rollback to
v1.61.0.
The minimum supported Rust version has been increased from v1.60.0 to v1.61.0.
Users building from source will need to ensure their rustc version is up to
date.
Synapse v1.81.0 deprecated application service authorization via query parameters as this is
considered insecure - and from Synapse v1.71.0 forwards the application service token has also been sent via
the Authorization header], making the insecure
query parameter authorization redundant. Since removing the ability to continue to use query parameters could break
backwards compatibility it has now been put behind a configuration option, use_appservice_legacy_authorization.
This option defaults to false, but can be activated by adding
Application services can no longer call /register with a user property to create new users.
The standard username property should be used instead. See the
Application Service specification
for more information.
Please ensure that you have migrated to using main on your shared configuration's instance_map
(or create one if necessary). This is required if you have any workers at all;
administrators of single-process (monolith) installations don't need to do anything.
The minimum supported Rust version has been increased from v1.58.1 to v1.60.0.
Users building from source will need to ensure their rustc version is up to
date.
Application services should ensure they call the /register endpoint with a
username property. The legacy user property is considered deprecated and
should no longer be included.
A future version of Synapse (v1.88.0 or later) will remove support for legacy
application service login.
should now be removed from individual worker YAML configurations and the main process should instead be added to the instance_map
in the shared YAML configuration, using the name main.
The old worker_replication_* settings are now considered deprecated and are expected to be removed in Synapse v1.88.0.
Synapse now attempts the versioned appservice paths before falling back to the
legacy paths.
Usage of the legacy routes should be considered deprecated.
Additionally, Synapse has supported sending the application service access token
via the Authorization header
since v1.70.0. For backwards compatibility it is also sent as the access_token
query parameter. This is insecure and should be considered deprecated.
A future version of Synapse (v1.88.0 or later) will remove support for legacy
application service routes and query parameter authorization.
Before this update, the
POST /_matrix/client/v3/rooms/{roomId}/report/{eventId}
endpoint would return a 403 if a user attempted to report an event that they did not have access to.
This endpoint will now return a 404 in this case instead.
Clients that implement event reporting should check that their error handling code will handle this
change.
Synapse v1.79.0 deprecates the
on_threepid_bind
"third-party rules" Synapse module callback method in favour of a new module method,
on_add_user_third_party_identifier.
on_threepid_bind will be removed in a future version of Synapse. You should check whether any Synapse
modules in use in your deployment are making use of on_threepid_bind, and update them where possible.
The arguments and functionality of the new method are the same.
The justification behind the name change is that the old method's name, on_threepid_bind, was
misleading. A user is considered to "bind" their third-party ID to their Matrix ID only if they
do so via an identity server
(so that users on other homeservers may find them). But this method was not called in that case -
it was only called when a user added a third-party identifier on the local homeserver.
Module developers may also be interested in the related
on_remove_user_third_party_identifier
module callback method that was also added in Synapse v1.79.0. This new method is called when a
user removes a third-party identifier from their account.
Synapse 1.78.0 replaces the /_synapse/admin/v1/media/<server_name>/delete
admin API with an identical endpoint at /_synapse/admin/v1/media/delete. Please
update your tooling to use the new endpoint. The deprecated version will be removed
in a future release.
When joining a room for the first time, Synapse 1.76.0 will request a partial join from the other server by default. Previously, server admins had to opt-in to this using an experimental config flag.
Server admins can opt out of this feature for the time being by setting
Synapse has changed the format of the account data and devices replication
streams (between workers). This is a forwards- and backwards-incompatible
change: v1.75 workers cannot process account data replicated by v1.76 workers,
and vice versa.
Once all workers are upgraded to v1.76 (or downgraded to v1.75), account data
and device replication will resume as normal.
The minimum supported version of Poetry is now 1.3.2 (previously 1.2.0, since
Synapse 1.67). If you have used poetry install to
install Synapse from a source checkout, you should upgrade poetry: see its
installation instructions.
For all other installation methods, no acction is required.
If you want to take advantage of this feature you need to install PyICU,
the ICU native dependency and its development headers
so that PyICU can build since no prebuilt wheels are available.
You can follow the PyICU documentation to do so,
and then do pip install matrix-synapse[user-search] for a PyPI install.
Docker images and Debian packages need nothing specific as they already
include or specify ICU as an explicit dependency.
Synapse 1.74 queues a background update
to rebuild the user directory,
in order to fix missing or erroneous entries.
When this update begins, the user directory will be cleared out and rebuilt from
scratch. User directory lookups will be incomplete until the rebuild completes.
Admins can monitor the rebuild's progress by using the
Background update Admin API.
Synapse v1.69.0 included the deprecation of legacy Prometheus metric names
and offered an option to disable them.
Synapse v1.71.0 disabled legacy Prometheus metric names by default.
This version, v1.73.0, removes those legacy Prometheus metric names entirely.
This also means that the enable_legacy_metrics configuration option has been
removed; it will no longer be possible to re-enable the legacy metric names.
If you use metrics and have not yet updated your Grafana dashboard(s),
Prometheus console(s) or alerting rule(s), please consider doing so when upgrading
to this version.
Note that the included Grafana dashboard was updated in v1.72.0 to correct some
metric names which were missed when legacy metrics were disabled by default.
To align with spec (changed in
MSC3905), Synapse now
only considers local users to be interesting. In other words, the users namespace
regex is only be applied against local users of the homeserver.
Please note, this probably doesn't affect the expected behavior of your application
service, since an interesting local user in a room still means all messages in the room
(from local or remote users) will still be considered interesting. And matching a room
with the rooms or aliases namespace regex will still consider all events sent in the
room to be interesting to the application service.
If one of your application service's users regex was intending to match a remote user,
this will no longer match as you expect. The behavioral mismatch between matching all
local users and some remote users is why the spec was changed/clarified and this
caveat is no longer supported.
Synapse v1.71.0 disables legacy Prometheus metric names by default.
For administrators that still rely on them and have not yet had chance to update their
uses of the metrics, it's still possible to specify enable_legacy_metrics: true in
the configuration to re-enable them temporarily.
Synapse v1.73.0 will remove legacy metric names altogether and at that point,
it will no longer be possible to re-enable them.
If you do not use metrics or you have already updated your Grafana dashboard(s),
Prometheus console(s) and alerting rule(s), there is no action needed.
Synapse now includes information indicating if a receipt applies to a thread when
replicating it to other workers. This is a forwards- and backwards-incompatible
change: v1.68 and workers cannot process receipts replicated by v1.69 workers, and
vice versa.
Once all workers are upgraded to v1.69 (or downgraded to v1.68), receipts
replication will resume as normal.
In current versions of Synapse, some Prometheus metrics are emitted under two different names,
with one of the names being older but non-compliant with OpenMetrics and Prometheus conventions
and one of the names being newer but compliant.
Synapse v1.71.0 will turn the old metric names off by default.
For administrators that still rely on them and have not had chance to update their
uses of the metrics, it's possible to specify enable_legacy_metrics: true in
the configuration to re-enable them temporarily.
Synapse v1.73.0 will remove legacy metric names altogether and it will no longer
be possible to re-enable them.
The Grafana dashboard, Prometheus recording rules and Prometheus Consoles included
in the contrib directory in the Synapse repository have been updated to no longer
rely on the legacy names. These can be used on a current version of Synapse
because current versions of Synapse emit both old and new names.
You may need to update your alerting rules or any other rules that depend on
the names of Prometheus metrics.
If you want to test your changes before legacy names are disabled by default,
you may specify enable_legacy_metrics: false in your homeserver configuration.
Redis support was added in v1.13.0 with it becoming the recommended method in
v1.18.0. It replaced the old direct TCP connections (which was deprecated as of
v1.18.0) to the main process. With Redis, rather than all the workers connecting
to the main process, all the workers and the main process connect to Redis,
which relays replication commands between processes. This can give a significant
CPU saving on the main process and is a prerequisite for upcoming
performance improvements.
To migrate to Redis add the redis config,
and remove the TCP replication listener from config of the master and
worker_replication_port from worker config. Note that a HTTP listener with a
replication resource is still required.
From the next major release (v1.68.0) installing Synapse from a source checkout
will require a recent Rust compiler. Those using packages or
pip install matrix-synapse will not be affected.
The simplest way of installing Rust is via rustup.rs
From the next major release (v1.68.0) Synapse will require SQLite 3.27.0 or
higher. Synapse v1.67.0 will be the last major release supporting SQLite
versions 3.22 to 3.26.
Those using Docker images or Debian packages from Matrix.org will not be
affected. If you have installed from source, you should check the version of
SQLite used by Python with:
As of this version, Synapse no longer allows the tasks of verifying email address
ownership, and password reset confirmation, to be delegated to an identity server.
This removal was previously planned for Synapse 1.64.0, but was
delayed until now to give
homeserver administrators more notice of the change.
To continue to allow users to add email addresses to their homeserver accounts,
and perform password resets, make sure that Synapse is configured with a working
email server in the email configuration
section
(including, at a minimum, a notif_from setting.)
Specifying an email setting under account_threepid_delegates will now cause
an error at startup.
Synapse v1.66.0 will remove the ability to delegate the tasks of verifying email address ownership, and password reset confirmation, to an identity server.
If you require your homeserver to verify e-mail addresses or to support password resets via e-mail, please configure your homeserver with SMTP access so that it can send e-mails on its own behalf.
Consult the configuration documentation for more information.
The option that will be removed is account_threepid_delegates.email.
Synapse now includes a flag indicating if an event is an outlier when
replicating it to other workers. This is a forwards- and backwards-incompatible
change: v1.63 and workers cannot process events replicated by v1.64 workers, and
vice versa.
Once all workers are upgraded to v1.64 (or downgraded to v1.63), event
replication will resume as normal.
frozendict 2.3.3
has recently been released, which fixes a memory leak that occurs during /sync
requests. We advise server administrators who installed Synapse via pip to upgrade
frozendict with pip install --upgrade frozendict. The Docker image
matrixdotorg/synapse and the Debian packages from packages.matrix.org already
include the updated library.
As a followup to changes in v1.60.0, the following spam-checker callbacks have changed signature:
user_may_join_room
user_may_invite
user_may_send_3pid_invite
user_may_create_room
user_may_create_room_alias
user_may_publish_room
check_media_file_for_spam
For each of these methods, the previous callback signature has been deprecated.
Whereas callbacks used to return bool, they should now return Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes"].
For instance, if your module implements user_may_join_room as follows:
async def user_may_join_room(self, user_id: str, room_id: str, is_invited: bool)
if ...:
# Request is spam
return False
# Request is not spam
return True
you should rewrite it as follows:
async def user_may_join_room(self, user_id: str, room_id: str, is_invited: bool)
if ...:
# Request is spam, mark it as forbidden (you may use some more precise error
# code if it is useful).
return synapse.module_api.errors.Codes.FORBIDDEN
# Request is not spam, mark it as such.
return synapse.module_api.NOT_SPAM
This release of Synapse will add a unique index to the state_group_edges table, in order
to prevent accidentally introducing duplicate information (for example, because a database
backup was restored multiple times).
Duplicate rows being present in this table could cause drastic performance problems; see
issue 11779 for more details.
If your Synapse database already has had duplicate rows introduced into this table,
this could fail, with either of these errors:
On Postgres:
synapse.storage.background_updates - 623 - INFO - background_updates-0 - Adding index state_group_edges_unique_idx to state_group_edges
synapse.storage.background_updates - 282 - ERROR - background_updates-0 - Error doing update
...
psycopg2.errors.UniqueViolation: could not create unique index "state_group_edges_unique_idx"
DETAIL: Key (state_group, prev_state_group)=(2, 1) is duplicated.
BEGIN;
DELETE FROM state_group_edges WHERE (ctid, state_group, prev_state_group) IN (
SELECT row_id, state_group, prev_state_group
FROM (
SELECT
ctid AS row_id,
MIN(ctid) OVER (PARTITION BY state_group, prev_state_group) AS min_row_id,
state_group,
prev_state_group
FROM state_group_edges
) AS t1
WHERE row_id <> min_row_id
);
COMMIT;
At the command-line, use sqlite3 path/to/your-homeserver-database.db:
BEGIN;
DELETE FROM state_group_edges WHERE (rowid, state_group, prev_state_group) IN (
SELECT row_id, state_group, prev_state_group
FROM (
SELECT
rowid AS row_id,
MIN(rowid) OVER (PARTITION BY state_group, prev_state_group) AS min_row_id,
state_group,
prev_state_group
FROM state_group_edges
)
WHERE row_id <> min_row_id
);
COMMIT;
Whereas check_event_for_spam callbacks used to return Union[str, bool], they should now return Union["synapse.module_api.NOT_SPAM", "synapse.module_api.errors.Codes"].
This is part of an ongoing refactoring of the SpamChecker API to make it less ambiguous and more powerful.
If your module implements check_event_for_spam as follows:
async def check_event_for_spam(event):
if ...:
# Event is spam
return True
# Event is not spam
return False
you should rewrite it as follows:
async def check_event_for_spam(event):
if ...:
# Event is spam, mark it as forbidden (you may use some more precise error
# code if it is useful).
return synapse.module_api.errors.Codes.FORBIDDEN
# Event is not spam, mark it as such.
return synapse.module_api.NOT_SPAM
The names of user devices are no longer visible to users on other homeservers by default.
Device IDs are unaffected, as these are necessary to facilitate end-to-end encryption.
The synapse.app.appservice worker application type allowed you to configure a
single worker to use to notify application services of new events, as long
as this functionality was disabled on the main process with notify_appservices: False.
Further, the synapse.app.user_dir worker application type allowed you to configure
a single worker to be responsible for updating the user directory, as long as this
was disabled on the main process with update_user_directory: False.
To unify Synapse's worker types, the synapse.app.appservice worker application
type and the notify_appservices configuration option have been deprecated.
The synapse.app.user_dir worker application type and update_user_directory
configuration option have also been deprecated.
To get the same functionality as was provided by the deprecated options, it's now recommended that the synapse.app.generic_worker
worker application type is used and that the notify_appservices_from_worker and/or
update_user_directory_from_worker options are set to the name of a worker.
For the time being, the old options can be used alongside the new options to make
it easier to transition between the two configurations, however please note that:
the options must not contradict each other (otherwise Synapse won't start); and
the notify_appservices and update_user_directory options will be removed in a future release of Synapse.
Synapse v1.57.0 includes a change to the
way transaction IDs are managed for application services. If your deployment uses a dedicated
worker for application service traffic, it must be stopped when the database is upgraded
(which normally happens when the main process is upgraded), to ensure the change is made safely
without any risk of reusing transaction IDs.
Deployments which do not use separate worker processes can be upgraded as normal. Similarly,
deployments where no application services are in use can be upgraded as normal.
Recovering from an incorrect upgrade
If the database schema is upgraded without stopping the worker responsible
for AS traffic, then the following error may be given when attempting to start
a Synapse worker or master process:
**********************************************************************************
Error during initialisation:
Postgres sequence 'application_services_txn_id_seq' is inconsistent with associated
table 'application_services_txns'. This can happen if Synapse has been downgraded and
then upgraded again, or due to a bad migration.
To fix this error, shut down Synapse (including any and all workers)
and run the following SQL:
SELECT setval('application_services_txn_id_seq', (
SELECT GREATEST(MAX(txn_id), 0) FROM application_services_txns
));
See docs/postgres.md for more information.
There may be more information in the logs.
**********************************************************************************
This error may also be seen if Synapse is downgraded to an earlier version,
and then upgraded again to v1.57.0 or later.
In either case:
Ensure that the worker responsible for AS traffic is stopped.
Run the SQL command given in the error message via psql.
Synapse will refuse to start if registration is enabled without email, captcha, or token-based verification unless the new config
flag enable_registration_without_verification is set to "true".
Synapse now refuses to start when using PostgreSQL with non-C values for COLLATE and
CTYPE unless the config flag allow_unsafe_locale, found in the database section of
the configuration file, is set to true. See the PostgreSQL documentation
for more information and instructions on how to fix a database with incorrect values.
The synctl script
has been made an
entry point
and no longer exists at the root of Synapse's source tree. If you wish to use
synctl to manage your homeserver, you should invoke synctl directly, e.g.
synctl start instead of ./synctl start or /path/to/synctl start.
When installing from a wheel, sdist, or PyPI, a synctl executable is added
to your Python installation's bin. This should be on your PATH
automatically, though you might need to activate a virtual environment
depending on how you installed Synapse.
Synapse v1.55.0 drops support for Mjolnir 1.3.1 and earlier.
If you use the Mjolnir module to moderate your homeserver,
please upgrade Mjolnir to version 1.3.2 or later before upgrading Synapse.
This release removes support for the structured: true logging configuration
which was deprecated in Synapse v1.23.0. If your logging configuration contains
structured: true then it should be modified based on the
structured logging documentation.
Per the deprecation notice in Synapse v1.51.0, listeners of type webclient
are no longer supported and configuring them is a now a configuration error.
Configuring a non-HTTP(S) web_client_location configuration is is now a
configuration error. Since the webclient listener is no longer supported, this
setting only applies to the root path / of Synapse's web server and no longer
the /_matrix/client/ path.
Formerly, entries in the cache were not evicted regardless of whether they were accessed after storing.
This behavior has now changed. By default entries in the cache are now evicted after 30m of not being accessed.
To change the default behavior, go to the caches section of the config and change the expire_caches and
cache_entry_ttl flags as necessary. Please note that these flags replace the expiry_time flag in the config.
The expiry_time flag will still continue to work, but it has been deprecated and will be removed in the future.
As announced with the release of Synapse 1.47.0,
the deprecated user_may_create_room_with_invites module callback has been removed.
Modules relying on it can instead implement user_may_invite
and use the get_room_state
module API to infer whether the invite is happening while creating a room (see this function
as an example). Alternately, modules can also implement on_create_room.
Note that Twisted 22.1.0
has recently been released, which fixes a security issue
within the Twisted library. We do not believe Synapse is affected by this vulnerability,
though we advise server administrators who installed Synapse via pip to upgrade Twisted
with pip install --upgrade Twisted treq as a matter of good practice. The Docker image
matrixdotorg/synapse and the Debian packages from packages.matrix.org are using the
updated library.
The user_may_create_room_with_invites is deprecated and will be removed in a future
version of Synapse. Modules implementing this callback can instead implement
user_may_invite
and use the get_room_state
module API method to infer whether the invite is happening in the context of creating a
room.
Media storage provider modules that read from the Synapse configuration object (i.e. that
read the value of hs.config.[...]) now need to specify the configuration section they're
reading from. This means that if a module reads the value of e.g. hs.config.media_store_path,
it needs to replace it with hs.config.media.media_store_path.
The url_cache/ and url_cache_thumbnails/ directories in the media store are
no longer mirrored to storage providers. These two directories can be safely
deleted from any configured storage providers to reclaim space.
The available worker applications documentation
has been updated to reflect that calls to the /spaces, /hierarchy, and
/summary endpoints can now be routed to workers for both client API and
federation requests.
Since Synapse 1.6.0 (2019-11-26) you can set a proxy for outbound HTTP requests via
http_proxy/https_proxy environment variables. This proxy was set for:
push
url previews
phone-home stats
recaptcha validation
CAS auth validation
OpenID Connect
Federation (checking public key revocation)
In this version we have added support for outbound requests for:
Outbound federation
Downloading remote media
Fetching public keys of other servers
These requests use the same proxy configuration. If you have a proxy configuration we
recommend to verify the configuration. It may be necessary to adjust the no_proxy
environment variable.
The template_dir settings in the sso, account_validity and email sections of the
configuration file are now deprecated. Server admins should use the new
templates.custom_template_directory setting in the configuration file and use one single
custom template directory for all aforementioned features. Template file names remain
unchanged. See the related documentation
for more information and examples.
We plan to remove support for these settings in October 2021.
The media repository worker documentation
has been updated to reflect that calls to /_synapse/admin/v1/users/{userId}/media
must now be handled by media repository workers. This is due to the new DELETE method
of this endpoint modifying the media store.
The current third-party rules module interface is deprecated in favour of the new generic
modules system introduced in Synapse v1.37.0. Authors of third-party rules modules can refer
to this documentation
to update their modules. Synapse administrators can refer to this documentation
to update their configuration once the modules they are using have been updated.
We plan to remove support for the current third-party rules interface in September 2021.
This release includes a database schema update which requires re-indexing one of
the larger tables in the database, events. This could result in increased
disk I/O for several hours or days after upgrading while the migration
completes. Furthermore, because we have to keep the old indexes until the new
indexes are ready, it could result in a significant, temporary, increase in
disk space.
To get a rough idea of the disk space required, check the current size of one
of the indexes. For example, from a psql shell, run the following sql:
We need to rebuild four indexes, so you will need to multiply this result
by four to give an estimate of the disk space required. For example, on one
particular server:
The current spam checker interface is deprecated in favour of a new generic modules system.
Authors of spam checker modules can refer to [this
documentation](modules/porting_legacy_module.md
to update their modules. Synapse administrators can refer to this
documentation
to update their configuration once the modules they are using have been updated.
We plan to remove support for the current spam checker interface in August 2021.
More module interfaces will be ported over to this new generic system in future versions
of Synapse.
This may affect you if you have enabled the account validity feature,
and have made use of a custom HTML template specified by the
account_validity.template_dir or
account_validity.account_renewed_html_path Synapse config options.
The template can now accept an expiration_ts variable, which
represents the unix timestamp in milliseconds for the future date of
which their account has been renewed until. See the default
template
for an example of usage.
ALso note that a new HTML template, account_previously_renewed.html,
has been added. This is is shown to users when they attempt to renew
their account with a valid renewal token that has already been used
before. The default template contents can been found
here,
and can also accept an expiration_ts variable. This template replaces
the error message users would previously see upon attempting to use a
valid renewal token more than once.
This release introduces a
regression that can
overwhelm connected Prometheus instances. This issue is not present in
Synapse v1.32.0rc1.
If you have been affected, please downgrade to 1.31.0. You then may need
to remove excess writeahead logs in order for Prometheus to recover.
Instructions for doing so are provided
here.
In compliance with the Application Service
spec,
Application Services are now required to use the
m.login.application_service type when registering users via the
/_matrix/client/r0/register endpoint. This behaviour was deprecated in
Synapse v1.30.0.
Please ensure your Application Services are up to date.
When using Synapse with a reverse proxy (in particular, when using the
x_forwarded option on an HTTP listener), Synapse now
expects to receive an X-Forwarded-Proto header on incoming
HTTP requests. If it is not set, Synapse will log a warning on each
received request.
To avoid the warning, administrators using a reverse proxy should ensure
that the reverse proxy sets X-Forwarded-Proto header to
https or http to indicate the protocol used
by the client.
Synapse also requires the Host header to be preserved.
See the reverse proxy documentation, where the
example configurations have been updated to show how to set these
headers.
(Users of Caddy are unaffected, since we
believe it sets X-Forwarded-Proto by default.)
This version changes the URI used for callbacks from OAuth2 and SAML2
identity providers:
If your server is configured for single sign-on via an OpenID
Connect or OAuth2 identity provider, you will need to add
[synapse public baseurl]/_synapse/client/oidc/callback to the list
of permitted "redirect URIs" at the identity provider.
See the OpenID docs for more information on setting
up OpenID Connect.
If your server is configured for single sign-on via a SAML2 identity
provider, you will need to add
[synapse public baseurl]/_synapse/client/saml2/authn_response as a
permitted "ACS location" (also known as "allowed callback URLs")
at the identity provider.
The "Issuer" in the "AuthnRequest" to the SAML2 identity
provider is also updated to
[synapse public baseurl]/_synapse/client/saml2/metadata.xml. If
your SAML2 identity provider uses this property to validate or
otherwise identify Synapse, its configuration will need to be
updated to use the new URL. Alternatively you could create a new,
separate "EntityDescriptor" in your SAML2 identity provider with
the new URLs and leave the URLs in the existing "EntityDescriptor"
as they were.
The HTML templates for SSO and email notifications now have Jinja2's
autoescape
enabled for files ending in .html, .htm, and .xml. If you have
customised these templates and see issues when viewing them you might
need to update them. It is expected that most configurations will need
no changes.
If you have customised the templates names for these templates, it is
recommended to verify they end in .html to ensure autoescape is
enabled.
The above applies to the following templates:
add_threepid.html
add_threepid_failure.html
add_threepid_success.html
notice_expiry.html
notice_expiry.html
notif_mail.html (which, by default, includes room.html and
notif.html)
v1.26.0 includes a lot of large changes. If something problematic
occurs, you may want to roll-back to a previous version of Synapse.
Because v1.26.0 also includes a new database schema version, reverting
that version is also required alongside the generic rollback
instructions mentioned above. In short, to roll back to v1.25.0 you need
to:
Stop the server
Decrease the schema version in the database:
UPDATE schema_version SET version = 58;
Delete the ignored users & chain cover data:
DROP TABLE IF EXISTS ignored_users;
UPDATE rooms SET has_auth_chain_index = false;
DELETE FROM event_auth_chain_links;
DELETE FROM event_auth_chains;
Mark the deltas as not run (so they will re-run on upgrade).
DELETE FROM applied_schema_deltas WHERE version = 59 AND file = "59/01ignored_user.py";
DELETE FROM applied_schema_deltas WHERE version = 59 AND file = "59/06chain_cover_index.sql";
Downgrade Synapse by following the instructions for your
installation method in the "Rolling back to older versions"
section above.
This is the last release of Synapse which guarantees support with Python
3.5, which passed its upstream End of Life date several months ago.
We will attempt to maintain support through March 2021, but without
guarantees.
In the future, Synapse will follow upstream schedules for ending support
of older versions of Python and PostgreSQL. Please upgrade to at least
Python 3.6 and PostgreSQL 9.6 as soon as possible.
Synapse v1.25.0 includes new settings, ip_range_blacklist and
ip_range_whitelist, for controlling outgoing requests from Synapse for
federation, identity servers, push, and for checking key validity for
third-party invite events. The previous setting,
federation_ip_range_blacklist, is deprecated. The new
ip_range_blacklist defaults to private IP ranges if it is not defined.
If you have never customised federation_ip_range_blacklist it is
recommended that you remove that setting.
If you have customised federation_ip_range_blacklist you should update
the setting name to ip_range_blacklist.
If you have a custom push server that is reached via private IP space
you may need to customise ip_range_blacklist or ip_range_whitelist.
This release allows the OpenID Connect mapping provider to perform
normalisation of the localpart of the Matrix ID. This allows for the
mapping provider to specify different algorithms, instead of the
default
way.
If your Synapse configuration uses a custom mapping provider
(oidc_config.user_mapping_provider.module is specified and
not equal to
synapse.handlers.oidc_handler.JinjaOidcMappingProvider)
then you must ensure that map_user_attributes of the
mapping provider performs some normalisation of the
localpart returned. To match previous behaviour you can
use the map_username_to_mxid_localpart function provided
by Synapse. An example is shown below:
from synapse.types import map_username_to_mxid_localpart
class MyMappingProvider:
def map_user_attributes(self, userinfo, token):
# ... your custom logic ...
sso_user_id = ...
localpart = map_username_to_mxid_localpart(sso_user_id)
return {"localpart": localpart}
The deprecation of the old endpoints was announced with Synapse 1.20.0
(released on 2020-09-22) and makes it easier for homeserver admins to
lock down external access to the Admin API endpoints.
This release deprecates use of the structured: true logging
configuration for structured logging. If your logging configuration
contains structured: true then it should be modified based on the
structured logging documentation.
The structured and drains logging options are now deprecated and
should be replaced by standard logging configuration of handlers and
formatters.
A future will release of Synapse will make using structured: true an
error.
This release introduces a backwards-incompatible change to modules
making use of ThirdPartyEventRules in Synapse. If you make use of a
module defined under the third_party_event_rules config option, please
make sure it is updated to handle the below change:
The http_client argument is no longer passed to modules as they are
initialised. Instead, modules are expected to make use of the
http_client property on the ModuleApi class. Modules are now passed
a module_api argument during initialisation, which is an instance of
ModuleApi. ModuleApi instances have a http_client property which
acts the same as the http_client argument previously passed to
ThirdPartyEventRules modules.
The reverse proxy documentation
has been updated to include reverse proxy directives for
/_synapse/client/* endpoints. As the user password reset flow now uses
endpoints under this prefix, you must update your reverse proxy
configurations for user password reset to work.
A new HTML template,
password_reset_confirmation.html,
has been added to the synapse/res/templates directory. If you are
using a custom template directory, you may want to copy the template
over and modify it.
Note that as of v1.20.0, templates do not need to be included in custom
template directories for Synapse to start. The default templates will be
used if a custom template cannot be found.
This page will appear to the user after clicking a password reset link
that has been emailed to them.
To complete password reset, the page must include a way to make a
POST request to
/_synapse/client/password_reset/{medium}/submit_token with the query
parameters from the original link, presented as a URL-encoded form. See
the file itself for more details.
The saml_error.html template was removed from Synapse and replaced
with the sso_error.html template. If your Synapse is configured to use
SAML and a custom sso_redirect_confirm_template_dir configuration then
any customisations of the saml_error.html template will need to be
merged into the sso_error.html template. These templates are similar,
but the parameters are slightly different:
The msg parameter should be renamed to error_description.
There is no longer a code parameter for the response code.
A string error parameter is available that includes a short hint
of why a user is seeing the error page.
From 10th August 2020, we will no longer publish Docker images with the
-py3 tag suffix. The images tagged with the
-py3 suffix have been identical to the non-suffixed tags
since release 0.99.0, and the suffix is obsolete.
On 10th August, we will remove the latest-py3 tag.
Existing per-release tags (such as v1.18.0-py3 will not
be removed, but no new -py3 tags will be added.
Scripts relying on the -py3 suffix will need to be
updated.
When setting up worker processes, we now recommend the use of a Redis
server for replication. The old direct TCP connection method is
deprecated and will be removed in a future release. See
the worker documentation for more details.
This version includes a database update which is run as part of the
upgrade, and which may take a couple of minutes in the case of a large
server. Synapse will not respond to HTTP requests while this update is
taking place.
A bug was introduced in Synapse 1.4.0 which could cause the room
directory to be incomplete or empty if Synapse was upgraded directly
from v1.2.1 or earlier, to versions between v1.4.0 and v1.12.x.
This will not be a problem for Synapse installations which were:
: - created at v1.4.0 or later,
- upgraded via v1.3.x, or
- upgraded straight from v1.2.1 or earlier to v1.13.0 or later.
If completeness of the room directory is a concern, installations which
are affected can be repaired as follows:
Run the following sql from a psql or
sqlite3 console:
INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES
('populate_stats_process_rooms', '{}', 'current_state_events_membership');
INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES
('populate_stats_process_users', '{}', 'populate_stats_process_rooms');
New templates (sso_auth_confirm.html, sso_auth_success.html, and
sso_account_deactivated.html) were added to Synapse. If your Synapse
is configured to use SSO and a custom
sso_redirect_confirm_template_dir configuration then these templates
will need to be copied from
synapse/res/templates into that directory.
Plugins using the complete_sso_login method of
synapse.module_api.ModuleApi should update to using the async/await
version complete_sso_login_async which includes additional checks. The
non-async version is considered deprecated.
v1.13.0 includes a lot of large changes. If something problematic
occurs, you may want to roll-back to a previous version of Synapse.
Because v1.13.0 also includes a new database schema version, reverting
that version is also required alongside the generic rollback
instructions mentioned above. In short, to roll back to v1.12.4 you need
to:
Stop the server
Decrease the schema version in the database:
UPDATE schema_version SET version = 57;
Downgrade Synapse by following the instructions for your
installation method in the "Rolling back to older versions"
section above.
This version includes a database update which is run as part of the
upgrade, and which may take some time (several hours in the case of a
large server). Synapse will not respond to HTTP requests while this
update is taking place.
This is only likely to be a problem in the case of a server which is
participating in many rooms.
As with all upgrades, it is recommended that you have a recent
backup of your database which can be used for recovery in the event
of any problems.
As an initial check to see if you will be affected, you can try
running the following query from the psql or
sqlite3 console. It is safe to run it while Synapse is
still running.
SELECT MAX(q.v) FROM (
SELECT (
SELECT ej.json AS v
FROM state_events se INNER JOIN event_json ej USING (event_id)
WHERE se.room_id=rooms.room_id AND se.type='m.room.create' AND se.state_key=''
LIMIT 1
) FROM rooms WHERE rooms.room_version IS NULL
) q;
This query will take about the same amount of time as the upgrade
process: ie, if it takes 5 minutes, then it is likely that Synapse
will be unresponsive for 5 minutes during the upgrade.
If you consider an outage of this duration to be acceptable, no
further action is necessary and you can simply start Synapse 1.12.0.
If you would prefer to reduce the downtime, continue with the steps
below.
The easiest workaround for this issue is to manually create a new
index before upgrading. On PostgreSQL, his can be done as follows:
CREATE INDEX CONCURRENTLY tmp_upgrade_1_12_0_index
ON state_events(room_id) WHERE type = 'm.room.create';
The above query may take some time, but is also safe to run while
Synapse is running.
We assume that no SQLite users have databases large enough to be
affected. If you are affected, you can run a similar query,
omitting the CONCURRENTLY keyword. Note however that this
operation may in itself cause Synapse to stop running for some time.
Synapse admins are reminded that SQLite is not recommended for use
outside a test environment.
Once the index has been created, the SELECT query in step 1 above
should complete quickly. It is therefore safe to upgrade to Synapse
1.12.0.
Once Synapse 1.12.0 has successfully started and is responding to
HTTP requests, the temporary index can be removed:
Specifying a log_file config option will now cause Synapse to refuse
to start, and should be replaced by with the log_config option.
Support for the log_file option was removed in v1.3.0 and has since
had no effect.
In an attempt to configure Synapse in a privacy preserving way, the
default behaviours of allow_public_rooms_without_auth and
allow_public_rooms_over_federation have been inverted. This means that
by default, only authenticated users querying the Client/Server API will
be able to query the room directory, and relatedly that the server will
not share room directory information with other servers over federation.
If your installation does not explicitly set these settings one way or
the other and you want either setting to be true then it will
necessary to update your homeserver configuration file accordingly.
For more details on the surrounding context see our
explainer.
This release includes a database migration which may take several
minutes to complete if there are a large number (more than a million or
so) of entries in the devices table. This is only likely to a be a
problem on very large installations.
If you have configured a custom template directory with the
email.template_dir option, be aware that there are new templates
regarding registration and threepid management (see below) that must be
included.
registration.html and registration.txt
registration_success.html and registration_failure.html
add_threepid.html and add_threepid.txt
add_threepid_failure.html and add_threepid_success.html
Synapse will expect these files to exist inside the configured template
directory, and will fail to start if they are absent. To view the
default templates, see
synapse/res/templates.
Note: As of this release, users will be unable to add phone numbers or
email addresses to their accounts, without changes to the Synapse
configuration. This includes adding an email address during
registration.
It is possible for a user to associate an email address or phone number
with their account, for a number of reasons:
for use when logging in, as an alternative to the user id.
in the case of email, as an alternative contact to help with account
recovery.
in the case of email, to receive notifications of missed messages.
Before an email address or phone number can be added to a user's
account, or before such an address is used to carry out a
password-reset, Synapse must confirm the operation with the owner of the
email address or phone number. It does this by sending an email or text
giving the user a link or token to confirm receipt. This process is
known as '3pid verification'. ('3pid', or 'threepid', stands for
third-party identifier, and we use it to refer to external identifiers
such as email addresses and phone numbers.)
Previous versions of Synapse delegated the task of 3pid verification to
an identity server by default. In most cases this server is vector.im
or matrix.org.
In Synapse 1.4.0, for security and privacy reasons, the homeserver will
no longer delegate this task to an identity server by default. Instead,
the server administrator will need to explicitly decide how they would
like the verification messages to be sent.
In the medium term, the vector.im and matrix.org identity servers
will disable support for delegated 3pid verification entirely. However,
in order to ease the transition, they will retain the capability for a
limited period. Delegated email verification will be disabled on Monday
2nd December 2019 (giving roughly 2 months notice). Disabling delegated
SMS verification will follow some time after that once SMS verification
support lands in Synapse.
Once delegated 3pid verification support has been disabled in the
vector.im and matrix.org identity servers, all Synapse versions that
depend on those instances will be unable to verify email and phone
numbers through them. There are no imminent plans to remove delegated
3pid verification from Sydent generally. (Sydent is the identity server
project that backs the vector.im and matrix.org instances).
To configure an SMTP server for Synapse, modify the configuration
section headed email, and be sure to have at least the
smtp_host, smtp_port and notif_from fields filled out.
You may also need to set smtp_user, smtp_pass, and
require_transport_security.
Some admins will wish to continue using email verification as part of
the registration process, but will not immediately have an appropriate
SMTP server at hand.
To this end, we will continue to support email verification delegation
via the vector.im and matrix.org identity servers for two months.
Support for delegated email verification will be disabled on Monday 2nd
December.
The account_threepid_delegates dictionary defines whether the
homeserver should delegate an external server (typically an identity
server) to handle
sending confirmation messages via email and SMS.
So to delegate email verification, in homeserver.yaml, set
account_threepid_delegates.email to the base URL of an identity
server. For example:
account_threepid_delegates:
email: https://example.com # Delegate email sending to example.com
Note that account_threepid_delegates.email replaces the deprecated
email.trust_identity_server_for_password_resets: if
email.trust_identity_server_for_password_resets is set to true, and
account_threepid_delegates.email is not set, then the first entry in
trusted_third_party_id_servers will be used as the
account_threepid_delegate for email. This is to ensure compatibility
with existing Synapse installs that set up external server handling for
these tasks before v1.4.0. If
email.trust_identity_server_for_password_resets is true and no
trusted identity server domains are configured, Synapse will report an
error and refuse to start.
If email.trust_identity_server_for_password_resets is false or
absent and no email delegate is configured in
account_threepid_delegates, then Synapse will send email verification
messages itself, using the configured SMTP server (see above). that
type.
Synapse does not support phone-number verification itself, so the only
way to maintain the ability for users to add phone numbers to their
accounts will be by continuing to delegate phone number verification to
the matrix.org and vector.im identity servers (or another identity
server that supports SMS sending).
The account_threepid_delegates dictionary defines whether the
homeserver should delegate an external server (typically an identity
server) to handle
sending confirmation messages via email and SMS.
So to delegate phone number verification, in homeserver.yaml, set
account_threepid_delegates.msisdn to the base URL of an identity
server. For example:
account_threepid_delegates:
msisdn: https://example.com # Delegate sms sending to example.com
The matrix.org and vector.im identity servers will continue to
support delegated phone number verification via SMS until such time as
it is possible for admins to configure their servers to perform phone
number verification directly. More details will follow in a future
release.
If you encounter problems with v1.4.0, it should be possible to roll
back to v1.3.1, subject to the following:
The 'room statistics' engine was heavily reworked in this release
(see #5971),
including significant changes to the database schema, which are not
easily reverted. This will cause the room statistics engine to stop
updating when you downgrade.
The room statistics are essentially unused in v1.3.1 (in future
versions of Synapse, they will be used to populate the room
directory), so there should be no loss of functionality. However,
the statistics engine will write errors to the logs, which can be
avoided by setting the following in homeserver.yaml:
stats:
enabled: false
Don't forget to re-enable it when you upgrade again, in preparation
for its use in the room directory!
Synapse v1.1.0 has a minimum Python requirement of Python 3.5. Python
3.6 or Python 3.7 are recommended as they have improved internal string
handling, significantly reducing memory usage.
If you use current versions of the Matrix.org-distributed Debian
packages or Docker images, action is not required.
If you install Synapse in a Python virtual environment, please see
"Upgrading to v0.34.0" for notes on setting up a new virtualenv under
Python 3.
If using PostgreSQL under Synapse, you will need to use PostgreSQL 9.5
or above. Please see the PostgreSQL
documentation for
more details on upgrading your database.
Synapse v1.0 is the first release to enforce validation of TLS
certificates for the federation API. It is therefore essential that your
certificates are correctly configured.
Note, v1.0 installations will also no longer be able to federate with
servers that have not correctly configured their certificates.
In rare cases, it may be desirable to disable certificate checking: for
example, it might be essential to be able to federate with a given
legacy server in a closed federation. This can be done in one of two
ways:-
Configure the global switch federation_verify_certificates to
false.
Configure a whitelist of server domains to trust via
federation_certificate_verification_whitelist.
When a user requests a password reset, Synapse will send an email to the
user to confirm the request.
Previous versions of Synapse delegated the job of sending this email to
an identity server. If the identity server was somehow malicious or
became compromised, it would be theoretically possible to hijack an
account through this means.
Therefore, by default, Synapse v1.0 will send the confirmation email
itself. If Synapse is not configured with an SMTP server, password reset
via email will be disabled.
To configure an SMTP server for Synapse, modify the configuration
section headed email, and be sure to have at least the smtp_host,
smtp_port and notif_from fields filled out. You may also need to set
smtp_user, smtp_pass, and require_transport_security.
If you are absolutely certain that you wish to continue using an
identity server for password resets, set
trust_identity_server_for_password_resets to true.
Some new templates have been added to the default template directory for the purpose of
the homeserver sending its own password reset emails. If you have configured a
custom template_dir in your Synapse config, these files will need to be added.
password_reset.html and password_reset.txt are HTML and plain text
templates respectively that contain the contents of what will be emailed
to the user upon attempting to reset their password via email.
password_reset_success.html and password_reset_failure.html are HTML
files that the content of which (assuming no redirect URL is set) will
be shown to the user after they attempt to click the link in the email
sent to them.
Please be aware that, before Synapse v1.0 is released around March 2019,
you will need to replace any self-signed certificates with those
verified by a root CA. Information on how to do so can be found at the
ACME docs.
This release is the first to fully support Python 3. Synapse will
now run on Python versions 3.5, or 3.6 (as well as 2.7). We
recommend switching to Python 3, as it has been shown to give
performance improvements.
For users who have installed Synapse into a virtualenv, we recommend
doing this by creating a new virtualenv. For example:
You can then start synapse as normal, having activated the new
virtualenv:
cd ~/synapse
source env3/bin/activate
synctl start
Users who have installed from distribution packages should see the
relevant package documentation. See below for notes on Debian
packages.
When upgrading to Python 3, you must make sure that your log
files are configured as UTF-8, by adding encoding: utf8 to the
RotatingFileHandler configuration (if you have one) in your
<server>.log.config file. For example, if your log.config
file contains:
There is no need to revert this change if downgrading to
Python 2.
We are also making available Debian packages which will run Synapse
on Python 3. You can switch to these packages with
apt-get install matrix-synapse-py3, however, please read
debian/NEWS
before doing so. The existing matrix-synapse packages will
continue to use Python 2 for the time being.
This release removes the riot.im from the default list of trusted
identity servers.
If riot.im is in your homeserver's list of
trusted_third_party_id_servers, you should remove it. It was added
in case a hypothetical future identity server was put there. If you
don't remove it, users may be unable to deactivate their accounts.
This release no longer installs the (unmaintained) Matrix Console
web client as part of the default installation. It is possible to
re-enable it by installing it separately and setting the
web_client_location config option, but please consider switching
to another client.
This release removes the example email notification templates from
res/templates (they are now internal to the python package). This
should only affect you if you (a) deploy your Synapse instance from a
git checkout or a github snapshot URL, and (b) have email notifications
enabled.
If you have email notifications enabled, you should ensure that
email.template_dir is either configured to point at a directory where
you have installed customised templates, or leave it unset to use the
default templates.
This release expands the anonymous usage stats sent if the opt-in
report_stats configuration is set to true. We now capture RSS memory
and cpu use at a very coarse level. This requires administrators to
install the optional psutil python module.
We would appreciate it if you could assist by ensuring this module is
available and report_stats is enabled. This will let us see if
performance changes to synapse are having an impact to the general
community.
If you want to use the new URL previewing API
(/_matrix/media/r0/preview_url) then you have to explicitly enable it
in the config and update your dependencies dependencies. See README.rst
for details.
This release includes the option to send anonymous usage stats to
matrix.org, and requires that administrators explicitly opt in or out by
setting the report_stats option to either true or false.
We would really appreciate it if you could help our project out by
reporting anonymized usage statistics from your homeserver. Only very
basic aggregate data (e.g. number of users) will be reported, but it
helps us to track the growth of the Matrix community, and helps us to
make Matrix a success, as well as to convince other networks that they
should peer with us.
Application services have had a breaking API change in this version.
They can no longer register themselves with a home server using the AS
HTTP API. This decision was made because a compromised application
service with free reign to register any regex in effect grants full
read/write access to the home server if a regex of .* is used. An
attack where a compromised AS re-registers itself with .* was deemed
too big of a security risk to ignore, and so the ability to register
with the HS remotely has been removed.
It has been replaced by specifying a list of application service
registrations in homeserver.yaml:
url: <String> # e.g. "https://my.application.service.com"
as_token: <String>
hs_token: <String>
sender_localpart: <String> # This is a new field which denotes the user_id localpart when using the AS token
namespaces:
users:
- exclusive: <Boolean>
regex: <String> # e.g. "@prefix_.*"
aliases:
- exclusive: <Boolean>
regex: <String>
rooms:
- exclusive: <Boolean>
regex: <String>
Where <db> is the location of the database,
<server_name> is the server name as specified in the
synapse configuration, and <signing_key> is the location
of the signing key as specified in the synapse configuration.
This may take some time to complete. Failures of signatures and content
hashes can safely be ignored.
Depending on precisely when you installed v0.5.0 you may have ended up
with a stale release of the reference matrix webclient installed as a
python module. To uninstall it and ensure you are depending on the
latest module, please run:
The webclient has been split out into a separate repository/package in
this release. Before you restart your homeserver you will need to pull
in the webclient package by running:
python setup.py develop --user
This release completely changes the database schema and so requires
upgrading it before starting the new version of the homeserver.
The script "database-prepare-for-0.5.0.sh" should be used to upgrade
the database. This will save all user information, such as logins and
profiles, but will otherwise purge the database. This includes messages,
which rooms the home server was a member of and room alias mappings.
If you would like to keep your history, please take a copy of your
database file and ask for help in #matrix:matrix.org. The upgrade
process is, unfortunately, non trivial and requires human intervention
to resolve any resulting conflicts during the upgrade process.
Before running the command the homeserver should be first completely
shutdown. To run it, simply specify the location of the database, e.g.:
Once this has successfully completed it will be safe to restart the
homeserver. You may notice that the homeserver takes a few seconds
longer to restart than usual as it reinitializes the database.
On startup of the new version, users can either rejoin remote rooms
using room aliases or by being reinvited. Alternatively, if any other
homeserver sends a message to a room that the homeserver was previously
in the local HS will automatically rejoin the room.
This registration API now closely matches the login API. This introduces
a bit more backwards and forwards between the HS and the client, but
this improves the overall flexibility of the API. You can now GET on
/register to retrieve a list of valid registration flows. Upon choosing
one, they are submitted in the same way as login, e.g:
{
type: m.login.password,
user: foo,
password: bar
}
The default HS supports 2 flows, with and without Identity Server email
authentication. Enabling captcha on the HS will add in an extra step to
all flows: m.login.recaptcha which must be completed before you can
transition to the next stage. There is a new login type:
m.login.email.identity which contains the threepidCreds key which
were previously sent in the original register request. For more
information on this, see the specification.
The VoIP specification has changed between v0.2.0 and v0.3.0. Users
should refresh any browser tabs to get the latest web client code. Users
on v0.2.0 of the web client will not be able to call those on v0.3.0 and
vice versa.
This release completely changes the database schema and so requires
upgrading it before starting the new version of the homeserver.
The script "database-prepare-for-0.0.1.sh" should be used to upgrade
the database. This will save all user information, such as logins and
profiles, but will otherwise purge the database. This includes messages,
which rooms the home server was a member of and room alias mappings.
Before running the command the homeserver should be first completely
shutdown. To run it, simply specify the location of the database, e.g.:
Once this has successfully completed it will be safe to restart the
homeserver. You may notice that the homeserver takes a few seconds
longer to restart than usual as it reinitializes the database.
On startup of the new version, users can either rejoin remote rooms
using room aliases or by being reinvited. Alternatively, if any other
homeserver sends a message to a room that the homeserver was previously
in the local HS will automatically rejoin the room.
Federation is the process by which users on different servers can participate
in the same room. For this to work, those other servers must be able to contact
yours to send messages.
The server_name configured in the Synapse configuration file (often
homeserver.yaml) defines how resources (users, rooms, etc.) will be
identified (eg: @user:example.com, #room:example.com). By default,
it is also the domain that other servers will use to try to reach your
server (via port 8448). This is easy to set up and will work provided
you set the server_name to match your machine's public DNS hostname.
For this default configuration to work, you will need to listen for TLS
connections on port 8448. The preferred way to do that is by using a
reverse proxy: see the reverse proxy documentation for instructions
on how to correctly set one up.
In some cases you might not want to run Synapse on the machine that has
the server_name as its public DNS hostname, or you might want federation
traffic to use a different port than 8448. For example, you might want to
have your user names look like @user:example.com, but you want to run
Synapse on synapse.example.com on port 443. This can be done using
delegation, which allows an admin to control where federation traffic should
be sent. See the delegation documentation for instructions on how to set this up.
Once federation has been configured, you should be able to join a room over
federation. A good place to start is #synapse:matrix.org - a room for
Synapse admins.
You can use the federation tester
to check if your homeserver is configured correctly. Alternatively try the
JSON API used by the federation tester.
Note that you'll have to modify this URL to replace DOMAIN with your
server_name. Hitting the API directly provides extra detail.
The typical failure mode for federation is that when the server tries to join
a room, it is rejected with "401: Unauthorized". Generally this means that other
servers in the room could not access yours. (Joining a room over federation is
a complicated dance which requires connections in both directions).
Another common problem is that people on other servers can't join rooms that
you invite them to. This can be caused by an incorrectly-configured reverse
proxy: see the reverse proxy documentation for instructions on how
to correctly configure a reverse proxy.
HTTP 308 Permanent Redirect redirects are not followed: Due to missing features
in the HTTP library used by Synapse, 308 redirects are currently not followed by
federating servers, which can cause M_UNKNOWN or 401 Unauthorized errors. This
may affect users who are redirecting apex-to-www (e.g. example.com -> www.example.com),
and especially users of the Kubernetes Nginx Ingress module, which uses 308 redirect
codes by default. For those Kubernetes users, this Stackoverflow post
might be helpful. For other users, switching to a 301 Moved Permanently code may be
an option. 308 redirect codes will be supported properly in a future
release of Synapse.
If you want to get up and running quickly with a trio of homeservers in a
private federation, there is a script in the demo directory. This is mainly
useful just for development purposes. See
demo scripts.
This section contains information on tweaking Synapse via the various options in the configuration file. A configuration
file should have been generated when you installed Synapse.
This is intended as a guide to the Synapse configuration. The behavior of a Synapse instance can be modified
through the many configuration settings documented here — each config option is explained,
including what the default is, how to change the default and what sort of behaviour the setting governs.
Also included is an example configuration for each setting. If you don't want to spend a lot of time
thinking about options, the config as generated sets sensible defaults for all values. Do note however that the
database defaults to SQLite, which is not recommended for production usage. You can read more on this subject
here.
The configuration file is a YAML file, which means that certain syntax rules
apply if you want your config file to be read properly. A few helpful things to know:
# before any option in the config will comment out that setting and either a default (if available) will
be applied or Synapse will ignore the setting. Thus, in example #1 below, the setting will be read and
applied, but in example #2 the setting will not be read and a default will be applied.
Example #1:
pid_file: DATADIR/homeserver.pid
Example #2:
#pid_file: DATADIR/homeserver.pid
Indentation matters! The indentation before a setting
will determine whether a given setting is read as part of another
setting, or considered on its own. Thus, in example #1, the enabled setting
is read as a sub-option of the presence setting, and will be properly applied.
However, the lack of indentation before the enabled setting in example #2 means
that when reading the config, Synapse will consider both presence and enabled as
different settings. In this case, presence has no value, and thus a default applied, and enabled
is an option that Synapse doesn't recognize and thus ignores.
Example #1:
presence:
enabled: false
Example #2:
presence:
enabled: false
In this manual, all top-level settings (ones with no indentation) are identified
at the beginning of their section (i.e. "### example_setting") and
the sub-options, if any, are identified and listed in the body of the section.
In addition, each setting has an example of its usage, with the proper indentation
shown.
Use the module sub-option to add modules under this option to extend functionality.
The module setting then has a sub-option, config, which can be used to define some configuration
for the module.
The server_name name will appear at the end of usernames and room addresses
created on your server. For example if the server_name was example.com,
usernames on your server would be in the format @user:example.com
In most cases you should avoid using a matrix specific subdomain such as
matrix.example.com or synapse.example.com as the server_name for the same
reasons you wouldn't use user@email.example.com as your email address.
See here
for information on how to host Synapse on a subdomain while preserving
a clean server_name.
The server_name cannot be changed later so it is important to
configure this correctly before you start Synapse. It should be all
lowercase and may contain an explicit port.
The public-facing base URL that clients use to access this Homeserver (not
including _matrix/...). This is the same URL a user might enter into the
'Custom Homeserver URL' field on their client. If you use Synapse with a
reverse proxy, this should be the URL to reach Synapse via the proxy.
Otherwise, it should be the URL to reach Synapse's client HTTP listener (see
'listeners' below).
By default, other servers will try to reach our server on port 8448, which can
be inconvenient in some environments.
Provided https://<server_name>/ on port 443 is routed to Synapse, this
option configures Synapse to serve a file at https://<server_name>/.well-known/matrix/server.
This will tell other servers to send traffic to port 443 instead.
This option allows server runners to add arbitrary key-value pairs to the client-facing .well-known response.
Note that the public_baseurl config option must be provided for Synapse to serve a response to /.well-known/matrix/client at all.
If this option is provided, it parses the given yaml to json and
serves it on /.well-known/matrix/client endpoint
alongside the standard properties.
Set the soft limit on the number of file descriptors synapse can use.
Zero is used to indicate synapse should set the soft limit to the hard limit.
Defaults to 0.
Presence tracking allows users to see the state (e.g online/offline)
of other local and remote users. Set the enabled sub-option to false to
disable presence tracking on this homeserver. Defaults to true.
This option replaces the previous top-level 'use_presence' option.
Example configuration:
presence:
enabled: false
enabled can also be set to a special value of "untracked" which ignores updates
received via clients and federation, while still accepting updates from the
module API.
The "untracked" option was added in Synapse 1.96.0.
Whether to require authentication to retrieve profile data (avatars, display names) of other
users through the client API. Defaults to false. Note that profile data is also available
via the federation API, unless allow_profile_lookup_over_federation is set to false.
Use this option to require a user to share a room with another user in order
to retrieve their profile information. Only checked on Client-Server
requests. Profile requests from other servers should be checked by the
requesting server. Defaults to false.
Use this option to prevent a user's profile data from being retrieved and
displayed in a room until they have joined it. By default, a user's
profile data is included in an invite event, regardless of the values
of the above two settings, and whether or not the users share a server.
Defaults to true.
If set to true, removes the need for authentication to access the server's
public rooms directory through the client API, meaning that anyone can
query the room directory. Defaults to false.
The minimum time in seconds between each GC for a generation, regardless of
the GC thresholds. This ensures that we don't do GC too frequently. A value of [1s, 10s, 30s]
indicates that a second must pass between consecutive generation 0 GCs, etc.
This option prevents outgoing requests from being sent to the specified blacklisted IP address
CIDR ranges. If this option is not specified then it defaults to private IP
address ranges (see the example below).
The blacklist applies to the outbound requests for federation, identity servers,
push servers, and for checking key validity for third-party invite events.
(0.0.0.0 and :: are always blacklisted, whether or not they are explicitly
listed here, since they correspond to unroutable addresses.)
This option replaces federation_ip_range_blacklist in Synapse v1.25.0.
Note: The value is ignored when an HTTP proxy is in use.
List of IP address CIDR ranges that should be allowed for federation,
identity servers, push servers, and for checking key validity for
third-party invite events. This is useful for specifying exceptions to
wide-ranging blacklisted target IP ranges - e.g. for communication with
a push server only visible in your network.
This whitelist overrides ip_range_blacklist and defaults to an empty
list.
List of ports that Synapse should listen on, their purpose and their
configuration.
Sub-options for each listener include:
port: the TCP port to bind to.
tag: An alias for the port in the logger name. If set the tag is logged instead
of the port. Default to None, is optional and only valid for listener with type: http.
See the docs request log format.
bind_addresses: a list of local addresses to listen on. The default is
'all local interfaces'.
type: the type of listener. Normally http, but other valid options are:
tls: set to true to enable TLS for this listener. Will use the TLS key/cert specified in tls_private_key_path / tls_certificate_path.
x_forwarded: Only valid for an 'http' listener. Set to true to use the X-Forwarded-For header as the client IP. Useful when Synapse is
behind a reverse-proxy.
request_id_header: The header extracted from each incoming request that is
used as the basis for the request ID. The request ID is used in
logs and tracing to
correlate and match up requests. When unset, Synapse will automatically
generate sequential request IDs. This option is useful when Synapse is behind
a reverse-proxy.
Added in Synapse 1.68.0.
resources: Only valid for an 'http' listener. A list of resources to host
on this port. Sub-options for each resource are:
names: a list of names of HTTP resources. See below for a list of valid resource names.
compress: set to true to enable gzip compression on HTTP bodies for this resource. This is currently only supported with the
client, consent, metrics and federation resources.
additional_resources: Only valid for an 'http' listener. A map of
additional endpoints which should be loaded via dynamic modules.
Unix socket support (Added in Synapse 1.89.0):
path: A path and filename for a Unix socket. Make sure it is located in a
directory with read and write permissions, and that it already exists (the directory
will not be created). Defaults to None.
Note: The use of both path and port options for the same listener is not
compatible.
The x_forwarded option defaults to true when using Unix sockets and can be omitted.
Other options that would not make sense to use with a UNIX socket, such as
bind_addresses and tls will be ignored and can be removed.
mode: The file permissions to set on the UNIX socket. Defaults to 666
Note: Must be set as type: http (does not support metrics and manhole).
Also make sure that metrics is not included in resources -> names
Valid resource names are:
client: the client-server API (/_matrix/client), and the synapse admin API (/_synapse/admin). Also implies media and static.
consent: user consent forms (/_matrix/consent). See here for more.
federation: the server-server API (/_matrix/federation). Also implies media, keys, openid
keys: the key discovery API (/_matrix/key).
media: the media API (/_matrix/media).
metrics: the metrics interface. See here. (Not compatible with Unix sockets)
replication: the HTTP replication API (/_synapse/replication). See here.
static: static resources under synapse/static (/_matrix/static). (Mostly useful for 'fallback authentication'.)
health: the health check endpoint. This endpoint
is by default active for all other resources and does not have to be activated separately.
This is only useful if you want to use the health endpoint explicitly on a dedicated port or
for workers and containers without listener e.g.
application services.
Example configuration #1:
listeners:
# TLS-enabled listener: for when matrix traffic is sent directly to synapse.
#
# (Note that you will also need to give Synapse a TLS key and certificate: see the TLS section
# below.)
#
- port: 8448
type: http
tls: true
resources:
- names: [client, federation]
Example configuration #2:
listeners:
# Insecure HTTP listener: for when matrix traffic passes through a reverse proxy
# that unwraps TLS.
#
# If you plan to use a reverse proxy, please see
# https://matrix-org.github.io/synapse/latest/reverse_proxy.html.
#
- port: 8008
tls: false
type: http
x_forwarded: true
bind_addresses: ['::1', '127.0.0.1']
resources:
- names: [client, federation]
compress: false
# example additional_resources:
additional_resources:
"/_matrix/my/custom/endpoint":
module: my_module.CustomRequestHandler
config: {}
# Turn on the twisted ssh manhole service on localhost on the given
# port.
- port: 9000
bind_addresses: ['::1', '127.0.0.1']
type: manhole
Example configuration #3:
listeners:
# Unix socket listener: Ideal for Synapse deployments behind a reverse proxy, offering
# lightweight interprocess communication without TCP/IP overhead, avoid port
# conflicts, and providing enhanced security through system file permissions.
#
# Note that x_forwarded will default to true, when using a UNIX socket. Please see
# https://matrix-org.github.io/synapse/latest/reverse_proxy.html.
#
- path: /run/synapse/main_public.sock
type: http
resources:
- names: [client, federation]
Connection settings for the manhole. You can find more information
on the manhole here. Manhole sub-options include:
username : the username for the manhole. This defaults to 'matrix'.
password: The password for the manhole. This defaults to 'rabbithole'.
ssh_priv_key_path and ssh_pub_key_path: The private and public SSH key pair used to encrypt the manhole traffic.
If these are left unset, then hardcoded and non-secret keys are used,
which could allow traffic to be intercepted if sent over a public network.
Forward extremities can build up in a room due to networking delays between
homeservers. Once this happens in a large room, calculation of the state of
that room can become quite expensive. To mitigate this, once the number of
forward extremities reaches a given threshold, Synapse will send an
org.matrix.dummy_event event, which will reduce the forward extremities
in the room.
This setting defines the threshold (i.e. number of forward extremities in the room) at which dummy events are sent.
The default value is 10.
An optional duration. If set, Synapse will run a daily background task to log out and
delete any device that hasn't been accessed for more than the specified amount of time.
Defaults to no duration, which means devices are never pruned.
Note: This task will always run on the main process, regardless of the value of
run_background_tasks_on. This is due to workers currently not having the ability to
delete devices.
Server admins can configure custom templates for email content. See
here for more information.
This setting has the following sub-options:
smtp_host: The hostname of the outgoing SMTP server to use. Defaults to 'localhost'.
smtp_port: The port on the mail server for outgoing SMTP. Defaults to 465 if force_tls is true, else 25.
Changed in Synapse 1.64.0: the default port is now aware of force_tls.
smtp_user and smtp_pass: Username/password for authentication to the SMTP server. By default, no
authentication is attempted.
force_tls: By default, Synapse connects over plain text and then optionally upgrades
to TLS via STARTTLS. If this option is set to true, TLS is used from the start (Implicit TLS),
and the option require_transport_security is ignored.
It is recommended to enable this if supported by your mail server.
New in Synapse 1.64.0.
require_transport_security: Set to true to require TLS transport security for SMTP.
By default, Synapse will connect over plain text, and will then switch to
TLS via STARTTLS if the SMTP server supports it. If this option is set,
Synapse will refuse to connect unless the server supports STARTTLS.
enable_tls: By default, if the server supports TLS, it will be used, and the server
must present a certificate that is valid for 'smtp_host'. If this option
is set to false, TLS will not be used.
notif_from: defines the "From" address to use when sending emails.
It must be set if email sending is enabled. The placeholder '%(app)s' will be replaced by the application name,
which is normally set in app_name, but may be overridden by the
Matrix client application. Note that the placeholder must be written '%(app)s', including the
trailing 's'.
app_name: app_name defines the default value for '%(app)s' in notif_from and email
subjects. It defaults to 'Matrix'.
enable_notifs: Set to true to enable sending emails for messages that the user
has missed. Disabled by default.
notif_for_new_users: Set to false to disable automatic subscription to email
notifications for new users. Enabled by default.
client_base_url: Custom URL for client links within the email notifications. By default
links will be based on "https://matrix.to". (This setting used to be called riot_base_url;
the old name is still supported for backwards-compatibility but is now deprecated.)
validation_token_lifetime: Configures the time that a validation email will expire after sending.
Defaults to 1h.
invite_client_location: The web client location to direct users to during an invite. This is passed
to the identity server as the org.matrix.web_client_location key. Defaults
to unset, giving no guidance to the identity server.
subjects: Subjects to use when sending emails from Synapse. The placeholder '%(app)s' will
be replaced with the value of the app_name setting, or by a value dictated by the Matrix client application.
In addition, each subject can use the following placeholders: '%(person)s', which will be replaced by the displayname
of the user(s) that sent the message(s), e.g. "Alice and Bob", and '%(room)s', which will be replaced by the name of the room the
message(s) have been sent to, e.g. "My super room". In addition, emails related to account administration will
can use the '%(server_name)s' placeholder, which will be replaced by the value of the
server_name setting in your Synapse configuration.
Here is a list of subjects for notification emails that can be set:
message_from_person_in_room: Subject to use to notify about one message from one or more user(s) in a
room which has a name. Defaults to "[%(app)s] You have a message on %(app)s from %(person)s in the %(room)s room..."
message_from_person: Subject to use to notify about one message from one or more user(s) in a
room which doesn't have a name. Defaults to "[%(app)s] You have a message on %(app)s from %(person)s..."
messages_from_person: Subject to use to notify about multiple messages from one or more users in
a room which doesn't have a name. Defaults to "[%(app)s] You have messages on %(app)s from %(person)s..."
messages_in_room: Subject to use to notify about multiple messages in a room which has a
name. Defaults to "[%(app)s] You have messages on %(app)s in the %(room)s room..."
messages_in_room_and_others: Subject to use to notify about multiple messages in multiple rooms.
Defaults to "[%(app)s] You have messages on %(app)s in the %(room)s room and others..."
messages_from_person_and_others: Subject to use to notify about multiple messages from multiple persons in
multiple rooms. This is similar to the setting above except it's used when
the room in which the notification was triggered has no name. Defaults to
"[%(app)s] You have messages on %(app)s from %(person)s and others..."
invite_from_person_to_room: Subject to use to notify about an invite to a room which has a name.
Defaults to "[%(app)s] %(person)s has invited you to join the %(room)s room on %(app)s..."
invite_from_person: Subject to use to notify about an invite to a room which doesn't have a
name. Defaults to "[%(app)s] %(person)s has invited you to chat on %(app)s..."
password_reset: Subject to use when sending a password reset email. Defaults to "[%(server_name)s] Password reset"
email_validation: Subject to use when sending a verification email to assert an address's
ownership. Defaults to "[%(server_name)s] Validate your email"
Example configuration:
email:
smtp_host: mail.server
smtp_port: 587
smtp_user: "exampleusername"
smtp_pass: "examplepassword"
force_tls: true
require_transport_security: true
enable_tls: false
notif_from: "Your Friendly %(app)s homeserver <noreply@example.com>"
app_name: my_branded_matrix_server
enable_notifs: true
notif_for_new_users: false
client_base_url: "http://localhost/riot"
validation_token_lifetime: 15m
invite_client_location: https://app.element.io
subjects:
message_from_person_in_room: "[%(app)s] You have a message on %(app)s from %(person)s in the %(room)s room..."
message_from_person: "[%(app)s] You have a message on %(app)s from %(person)s..."
messages_from_person: "[%(app)s] You have messages on %(app)s from %(person)s..."
messages_in_room: "[%(app)s] You have messages on %(app)s in the %(room)s room..."
messages_in_room_and_others: "[%(app)s] You have messages on %(app)s in the %(room)s room and others..."
messages_from_person_and_others: "[%(app)s] You have messages on %(app)s from %(person)s and others..."
invite_from_person_to_room: "[%(app)s] %(person)s has invited you to join the %(room)s room on %(app)s..."
invite_from_person: "[%(app)s] %(person)s has invited you to chat on %(app)s..."
password_reset: "[%(server_name)s] Password reset"
email_validation: "[%(server_name)s] Validate your email"
This option disables/enables monthly active user blocking. Used in cases where the admin or
server owner wants to limit to the number of monthly active users. When enabled and a limit is
reached the server returns a ResourceLimitError with error type Codes.RESOURCE_LIMIT_EXCEEDED.
Defaults to false. If this is enabled, a value for max_mau_value must also be set.
This option sets the hard limit of monthly active users above which the server will start
blocking user actions if limit_usage_by_mau is enabled. Defaults to 0.
The option mau_trial_days is a means to add a grace period for active users. It
means that users must be active for the specified number of days before they
can be considered active and guards against the case where lots of users
sign up in a short space of time never to return after their initial
session. Defaults to 0.
The option mau_appservice_trial_days is similar to mau_trial_days, but applies a different
trial number if the user was registered by an appservice. A value
of 0 means no trial days are applied. Appservices not listed in this dictionary
use the value of mau_trial_days instead.
The option mau_limit_alerting is a means of limiting client-side alerting
should the mau limit be reached. This is useful for small instances
where the admin has 5 mau seats (say) for 5 specific people and no
interest increasing the mau limit further. Defaults to true, which
means that alerting is enabled.
If enabled, the metrics for the number of monthly active users will
be populated, however no one will be limited based on these numbers. If limit_usage_by_mau
is true, this is implied to be true. Defaults to false.
Sometimes the server admin will want to ensure certain accounts are
never blocked by mau checking. These accounts are specified by this option.
Defaults to none. Add accounts by specifying the medium and address of the
reserved threepid (3rd party identifier).
When this option is enabled, the room "complexity" will be checked before a user
joins a new remote room. If it is above the complexity limit, the server will
disallow joining, or will instantly leave. This is useful for homeservers that are
resource-constrained. Options for this setting include:
enabled: whether this check is enabled. Defaults to false.
complexity: the limit above which rooms cannot be joined. The default is 1.0.
complexity_error: override the error which is returned when the room is too complex with a
custom message.
admins_can_join: allow server admins to join complex rooms. Default is false.
Room complexity is an arbitrary measure based on factors such as the number of
users in the room.
Example configuration:
limit_remote_rooms:
enabled: true
complexity: 0.5
complexity_error: "I can't let you do that, Dave."
admins_can_join: true
Whether to allow per-room membership profiles through the sending of membership
events with profile information that differs from the target's global profile.
Defaults to true.
How long to keep redacted events in unredacted form in the database. After
this period redacted events get replaced with their redacted form in the DB.
Synapse will check whether the rentention period has concluded for redacted
events every 5 minutes. Thus, even if this option is set to 0, Synapse may
still take up to 5 minutes to purge redacted events from the database.
Inhibits the /requestToken endpoints from returning an error that might leak
information about whether an e-mail address is in use or not on this
homeserver. Defaults to false.
Note that for some endpoints the error situation is the e-mail already being
used, and for others the error is entering the e-mail being unused.
If this option is enabled, instead of returning an error, these endpoints will
act as if no error happened and return a fake session ID ('sid') to clients.
A list of domains that the domain portion of next_link parameters
must match.
This parameter is optionally provided by clients while requesting
validation of an email or phone number, and maps to a link that
users will be automatically redirected to after validation
succeeds. Clients can make use this parameter to aid the validation
process.
The whitelist is applied whether the homeserver or an identity server is handling validation.
The default value is no whitelist functionality; all domains are
allowed. Setting this value to an empty list will instead disallow
all domains.
These options define templates to use when generating email or HTML page contents.
The custom_template_directory determines which directory Synapse will try to
find template files in to use to generate email or HTML page contents.
If not set, or a file is not found within the template directory, a default
template from within the Synapse package will be used.
See here for more
information about using custom templates.
This option and the associated options determine message retention policy at the
server level.
Room admins and mods can define a retention period for their rooms using the
m.room.retention state event, and server admins can cap this period by setting
the allowed_lifetime_min and allowed_lifetime_max config options.
If this feature is enabled, Synapse will regularly look for and purge events
which are older than the room's maximum retention period. Synapse will also
filter events received over federation so that events that should have been
purged are ignored and not stored again.
The message retention policies feature is disabled by default. You can read more
about this feature here.
This setting has the following sub-options:
default_policy: Default retention policy. If set, Synapse will apply it to rooms that lack the
'm.room.retention' state event. This option is further specified by the
min_lifetime and max_lifetime sub-options associated with it. Note that the
value of min_lifetime doesn't matter much because Synapse doesn't take it into account yet.
allowed_lifetime_min and allowed_lifetime_max: Retention policy limits. If
set, and the state of a room contains a m.room.retention event in its state
which contains a min_lifetime or a max_lifetime that's out of these bounds,
Synapse will cap the room's policy to these limits when running purge jobs.
purge_jobs and the associated shortest_max_lifetime and longest_max_lifetime sub-options:
Server admins can define the settings of the background jobs purging the
events whose lifetime has expired under the purge_jobs section.
If no configuration is provided for this option, a single job will be set up to delete
expired events in every room daily.
Each job's configuration defines which range of message lifetimes the job
takes care of. For example, if shortest_max_lifetime is '2d' and
longest_max_lifetime is '3d', the job will handle purging expired events in
rooms whose state defines a max_lifetime that's both higher than 2 days, and
lower than or equal to 3 days. Both the minimum and the maximum value of a
range are optional, e.g. a job with no shortest_max_lifetime and a
longest_max_lifetime of '3d' will handle every room with a retention policy
whose max_lifetime is lower than or equal to three days.
The rationale for this per-job configuration is that some rooms might have a
retention policy with a low max_lifetime, where history needs to be purged
of outdated messages on a more frequent basis than for the rest of the rooms
(e.g. every 12h), but not want that purge to be performed by a job that's
iterating over every room it knows, which could be heavy on the server.
If any purge job is configured, it is strongly recommended to have at least
a single job with neither shortest_max_lifetime nor longest_max_lifetime
set, or one job without shortest_max_lifetime and one job without
longest_max_lifetime set. Otherwise some rooms might be ignored, even if
allowed_lifetime_min and allowed_lifetime_max are set, because capping a
room's policy to these values is done after the policies are retrieved from
Synapse's database (which is done using the range specified in a purge job's
configuration).
This option specifies a PEM-encoded X509 certificate for TLS.
This certificate, as of Synapse 1.0, will need to be a valid and verifiable
certificate, signed by a recognised Certificate Authority. Defaults to none.
Be sure to use a .pem file that includes the full certificate chain including
any intermediate certificates (for instance, if using certbot, use
fullchain.pem as your certificate, not cert.pem).
The minimum TLS version that will be used for outbound federation requests.
Defaults to "1". Configurable to "1", "1.1", "1.2", or "1.3". Note
that setting this value higher than "1.2" will prevent federation to most
of the public Matrix network: only configure it to "1.3" if you have an
entirely private federation setup and you can ensure TLS 1.3 support.
Skip federation certificate verification on a given whitelist
of domains.
This setting should only be used in very specific cases, such as
federation over Tor hidden services and similar. For private networks
of homeservers, you likely want to use a private CA instead.
Only effective if federation_verify_certificates is true.
Restrict federation to the given whitelist of domains.
N.B. we recommend also firewalling your federation listener to limit
inbound federation traffic as early as possible, rather than relying
purely on this application-layer restriction. If not specified, the
default is to whitelist everything.
Note: this does not stop a server from joining rooms that servers not on the
whitelist are in. As such, this option is really only useful to establish a
"private federation", where a group of servers all whitelist each other and have
the same whitelist.
Report prometheus metrics on the age of PDUs being sent to and received from
the given domains. This can be used to give an idea of "delay" on inbound
and outbound federation, though be aware that any delay can be due to problems
at either end or with the intermediate network.
Set to false to disable profile lookup over federation. By default, the
Federation API allows other homeservers to obtain profile data of any user
on this homeserver.
Set this option to true to allow device display name lookup over federation. By default, the
Federation API prevents other homeservers from obtaining the display names of any user devices
on this homeserver.
The federation section defines some sub-options related to federation.
The following options are related to configuring timeout and retry logic for one request,
independently of the others.
Short retry algorithm is used when something or someone will wait for the request to have an
answer, while long retry is used for requests that happen in the background,
like sending a federation transaction.
client_timeout: timeout for the federation requests. Default to 60s.
max_short_retry_delay: maximum delay to be used for the short retry algo. Default to 2s.
max_long_retry_delay: maximum delay to be used for the short retry algo. Default to 60s.
max_short_retries: maximum number of retries for the short retry algo. Default to 3 attempts.
max_long_retries: maximum number of retries for the long retry algo. Default to 10 attempts.
The following options control the retry logic when communicating with a specific homeserver destination.
Unlike the previous configuration options, these values apply across all requests
for a given destination and the state of the backoff is stored in the database.
destination_min_retry_interval: the initial backoff, after the first request fails. Defaults to 10m.
destination_retry_multiplier: how much we multiply the backoff by after each subsequent fail. Defaults to 2.
destination_max_retry_interval: a cap on the backoff. Defaults to a week.
A cache 'factor' is a multiplier that can be applied to each of
Synapse's caches in order to increase or decrease the maximum
number of entries that can be stored.
caches can be configured through the following sub-options:
global_factor: Controls the global cache factor, which is the default cache factor
for all caches if a specific factor for that cache is not otherwise
set.
This can also be set by the SYNAPSE_CACHE_FACTOR environment
variable. Setting by environment variable takes priority over
setting through the config file.
Defaults to 0.5, which will halve the size of all caches.
per_cache_factors: A dictionary of cache name to cache factor for that individual
cache. Overrides the global cache factor for a given cache.
These can also be set through environment variables comprised
of SYNAPSE_CACHE_FACTOR_ + the name of the cache in capital
letters and underscores. Setting by environment variable
takes priority over setting through the config file.
Ex. SYNAPSE_CACHE_FACTOR_GET_USERS_WHO_SHARE_ROOM_WITH_USER=2.0
Some caches have '*' and other characters that are not
alphanumeric or underscores. These caches can be named with or
without the special characters stripped. For example, to specify
the cache factor for *stateGroupCache* via an environment
variable would be SYNAPSE_CACHE_FACTOR_STATEGROUPCACHE=2.0.
expire_caches: Controls whether cache entries are evicted after a specified time
period. Defaults to true. Set to false to disable this feature. Note that never expiring
caches may result in excessive memory usage.
cache_entry_ttl: If expire_caches is enabled, this flag controls how long an entry can
be in a cache without having been accessed before being evicted.
Defaults to 30m.
sync_response_cache_duration: Controls how long the results of a /sync request are
cached for after a successful response is returned. A higher duration can help clients
with intermittent connections, at the cost of higher memory usage.
A value of zero means that sync responses are not cached.
Defaults to 2m.
Changed in Synapse 1.62.0: The default was changed from 0 to 2m.
cache_autotuning and its sub-options max_cache_memory_usage, target_cache_memory_usage, and
min_cache_ttl work in conjunction with each other to maintain a balance between cache memory
usage and cache entry availability. You must be using jemalloc
to utilize this option, and all three of the options must be specified for this feature to work. This option
defaults to off, enable it by providing values for the sub-options listed below. Please note that the feature will not work
and may cause unstable behavior (such as excessive emptying of caches or exceptions) if all of the values are not provided.
Please see the Config Conventions for information on how to specify memory size and cache expiry
durations.
max_cache_memory_usage sets a ceiling on how much memory the cache can use before caches begin to be continuously evicted.
They will continue to be evicted until the memory usage drops below the target_memory_usage, set in
the setting below, or until the min_cache_ttl is hit. There is no default value for this option.
target_cache_memory_usage sets a rough target for the desired memory usage of the caches. There is no default value
for this option.
min_cache_ttl sets a limit under which newer cache entries are not evicted and is only applied when
caches are actively being evicted/max_cache_memory_usage has been exceeded. This is to protect hot caches
from being emptied while Synapse is evicting due to memory. There is no default value for this option.
The cache factors (i.e. caches.global_factor and caches.per_cache_factors) may be reloaded at any time by sending a
SIGHUP signal to Synapse using e.g.
kill -HUP [PID_OF_SYNAPSE_PROCESS]
If you are running multiple workers, you must individually update the worker
config file and send this signal to each worker process.
If you're using the example systemd service
file in Synapse's contrib directory, you can send a SIGHUP signal by using
systemctl reload matrix-synapse.
The database setting defines the database that synapse uses to store all of
its data.
Associated sub-options:
name: this option specifies the database engine to use: either sqlite3 (for SQLite)
or psycopg2 (for PostgreSQL). If no name is specified Synapse will default to SQLite.
txn_limit gives the maximum number of transactions to run per connection
before reconnecting. Defaults to 0, which means no limit.
allow_unsafe_locale is an option specific to Postgres. Under the default behavior, Synapse will refuse to
start if the postgres db is set to a non-C locale. You can override this behavior (which is not recommended)
by setting allow_unsafe_locale to true. Note that doing so may corrupt your database. You can find more information
here and here.
args gives options which are passed through to the database engine,
except for options starting with cp_, which are used to configure the Twisted
connection pool. For a reference to valid arguments, see:
The databases option allows specifying a mapping between certain database tables and
database host details, spreading the load of a single Synapse instance across multiple
database backends. This is often referred to as "database sharding". This option is only
supported for PostgreSQL database backends.
Important note: This is a supported option, but is not currently used in production by the
Matrix.org Foundation. Proceed with caution and always make backups.
databases is a dictionary of arbitrarily-named database entries. Each entry is equivalent
to the value of the database homeserver config option (see above), with the addition of
a data_stores key. data_stores is an array of strings that specifies the data store(s)
(a defined label for a set of tables) that should be stored on the associated database
backend entry.
The currently defined values for data_stores are:
"state": Database that relates to state groups will be stored in this database.
Specifically, that means the following tables:
state_groups
state_group_edges
state_groups_state
And the following sequences:
state_groups_seq_id
"main": All other database tables and sequences.
All databases will end up with additional tables used for tracking database schema migrations
and any pending background updates. Synapse will create these automatically on startup when checking for
and/or performing database schema migrations.
To migrate an existing database configuration (e.g. all tables on a single database) to a different
configuration (e.g. the "main" data store on one database, and "state" on another), do the following:
Take a backup of your existing database. Things can and do go wrong and database corruption is no joke!
Ensure all pending database migrations have been applied and background updates have run. The simplest
way to do this is to use the update_synapse_database script supplied with your Synapse installation.
Copy over the necessary tables and sequences from one database to the other. Tables relating to database
migrations, schemas, schema versions and background updates should not be copied.
As an example, say that you'd like to split out the "state" data store from an existing database which
currently contains all data stores.
Simply copy the tables and sequences defined above for the "state" datastore from the existing database
to the secondary database. As noted above, additional tables will be created in the secondary database
when Synapse is started.
Modify/create the databases option in your homeserver.yaml to match the desired database configuration.
Start Synapse. Check that it starts up successfully and that things generally seem to be working.
Drop the old tables that were copied in step 3.
Only one of the options database or databases may be specified in your config, but not both.
This is a ratelimiting option for messages that ratelimits sending based on the account the client
is using. It defaults to: per_second: 0.2, burst_count: 10.
This option checks the validity of registration tokens that ratelimits requests based on
the client's IP address.
Defaults to per_second: 0.1, burst_count: 5.
address ratelimits login requests based on the client's IP
address. Defaults to per_second: 0.003, burst_count: 5.
account ratelimits login requests based on the account the
client is attempting to log into. Defaults to per_second: 0.003,
burst_count: 5.
failed_attempts ratelimits login requests based on the account the
client is attempting to log into, based on the amount of failed login
attempts for this account. Defaults to per_second: 0.17, burst_count: 3.
This option sets ratelimiting redactions by room admins. If this is not explicitly
set then it uses the same ratelimiting as per rc_message. This is useful
to allow room admins to deal with abuse quickly.
This option allows for ratelimiting number of rooms a user can join. This setting has the following sub-options:
local: ratelimits when users are joining rooms the server is already in.
Defaults to per_second: 0.1, burst_count: 10.
remote: ratelimits when users are trying to join rooms not on the server (which
can be more computationally expensive than restricting locally). Defaults to
per_second: 0.01, burst_count: 10
This option allows admins to ratelimit joins to a room based on the number of recent
joins (local or remote) to that room. It is intended to mitigate mass-join spam
waves which target multiple homeservers.
By default, one join is permitted to a room every second, with an accumulating
buffer of up to ten instantaneous joins.
This option sets ratelimiting how often invites can be sent in a room or to a
specific user. per_room defaults to per_second: 0.3, burst_count: 10 and
per_user defaults to per_second: 0.003, burst_count: 5.
Client requests that invite user(s) when creating a
room
will count against the rc_invites.per_room limit, whereas
client requests to invite a single user to a
room
will count against both the rc_invites.per_user and rc_invites.per_room limits.
Federation requests to invite a user will count against the rc_invites.per_user
limit only, as Synapse presumes ratelimiting by room will be done by the sending server.
The rc_invites.per_user limit applies to the receiver of the invite, rather than the
sender, meaning that a rc_invite.per_user.burst_count of 5 mandates that a single user
cannot receive more than a burst of 5 invites at a time.
In contrast, the rc_invites.per_issuer limit applies to the issuer of the invite, meaning that a rc_invite.per_issuer.burst_count of 5 mandates that single user cannot send more than a burst of 5 invites at a time.
Changed in version 1.63: added the per_issuer limit.
This option ratelimits 3PID invites (i.e. invites sent to a third-party ID
such as an email address or a phone number) based on the account that's
sending the invite. Defaults to per_second: 0.2, burst_count: 10.
This option ratelimits creation of MXC URIs via the /_matrix/media/v1/create
endpoint based on the account that's creating the media. Defaults to
per_second: 10, burst_count: 50.
How many pending media uploads can a given user have? A pending media upload
is a created MXC URI that (a) is not expired (the unused_expires_at timestamp
has not passed) and (b) the media has not yet been uploaded for. Defaults to 5.
If you are using a reverse proxy you may also need to set this value in
your reverse proxy's config. Defaults to 50M. Notably Nginx has a small max body size by default.
See here for more on using a reverse proxy with Synapse.
A list of domains to never download media from. Media from these
domains that is already downloaded will not be deleted, but will be
inaccessible to users. This option does not affect admin APIs trying
to download/operate on media.
This will not prevent the listed domains from accessing media themselves.
It simply prevents users on this server from downloading media originating
from the listed servers.
This will have no effect on media originating from the local server.
This only affects media downloaded from other Matrix servers, to
block domains from URL previews see url_preview_url_blacklist.
Whether to generate new thumbnails on the fly to precisely match
the resolution requested by the client. If true then whenever
a new resolution is requested by the client the server will
generate a new thumbnail. If false the server will pick a thumbnail
from a precalculated list. Defaults to false.
Controls whether local media and entries in the remote media cache
(media that is downloaded from other homeservers) should be removed
under certain conditions, typically for the purpose of saving space.
Purging media files will be the carried out by the media worker
(that is, the worker that has the enable_media_repo homeserver config
option set to 'true'). This may be the main process.
The media_retention.local_media_lifetime and
media_retention.remote_media_lifetime config options control whether
media will be purged if it has not been accessed in a given amount of
time. Note that media is 'accessed' when loaded in a room in a client, or
otherwise downloaded by a local or remote user. If the media has never
been accessed, the media's creation time is used instead. Both thumbnails
and the original media will be removed. If either of these options are unset,
then media of that type will not be purged.
Local or cached remote media that has been
quarantined
will not be deleted. Similarly, local media that has been marked as
protected from quarantine
will not be deleted.
This setting determines whether the preview URL API is enabled.
It is disabled by default. Set to true to enable. If enabled you must specify a
url_preview_ip_range_blacklist blacklist.
List of IP address CIDR ranges that the URL preview spider is denied
from accessing. There are no defaults: you must explicitly
specify a list for URL previewing to work. You should specify any
internal services in your network that you do not want synapse to try
to connect to, otherwise anyone in any Matrix room could cause your
synapse to issue arbitrary GET requests to your internal services,
causing serious security issues.
(0.0.0.0 and :: are always blacklisted, whether or not they are explicitly
listed here, since they correspond to unroutable addresses.)
This must be specified if url_preview_enabled is set. It is recommended that
you use the following example list as a starting point.
Note: The value is ignored when an HTTP proxy is in use.
This option sets a list of IP address CIDR ranges that the URL preview spider is allowed
to access even if they are specified in url_preview_ip_range_blacklist.
This is useful for specifying exceptions to wide-ranging blacklisted
target IP ranges - e.g. for enabling URL previews for a specific private
website only visible in your network. Defaults to none.
Optional list of URL matches that the URL preview spider is
denied from accessing. You should use url_preview_ip_range_blacklist
in preference to this, otherwise someone could define a public DNS
entry that points to a private IP address and circumvent the blacklist.
This is more useful if you know there is an entire shape of URL that
you know that will never want synapse to try to spider.
Each list entry is a dictionary of url component attributes as returned
by urlparse.urlsplit as applied to the absolute form of the URL. See
here for more
information. Some examples are:
username
netloc
scheme
path
The values of the dictionary are treated as a filename match pattern
applied to that component of URLs, unless they start with a ^ in which
case they are treated as a regular expression match. If all the
specified component matches for a given list item succeed, the URL is
blacklisted.
Example configuration:
url_preview_url_blacklist:
# blacklist any URL with a username in its URI
- username: '*'
# blacklist all *.google.com URLs
- netloc: 'google.com'
- netloc: '*.google.com'
# blacklist all plain HTTP URLs
- scheme: 'http'
# blacklist http(s)://www.acme.com/foo
- netloc: 'www.acme.com'
path: '/foo'
# blacklist any URL with a literal IPv4 address
- netloc: '^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$'
A list of values for the Accept-Language HTTP header used when
downloading webpages during URL preview generation. This allows
Synapse to specify the preferred languages that URL previews should
be in when communicating with remote servers.
Each value is a IETF language tag; a 2-3 letter identifier for a
language, optionally followed by subtags separated by '-', specifying
a country or region variant.
Multiple values can be provided, and a weight can be added to each by
using quality value syntax (;q=). '*' translates to any language.
oEmbed allows for easier embedding content from a website. It can be
used for generating URLs previews of services which support it. A default list of oEmbed providers
is included with Synapse. Set disable_default_providers to true to disable using
these default oEmbed URLs. Use additional_providers to specify additional files with oEmbed configuration (each
should be in the form of providers.json). By default this list is empty.
Whether guests should be allowed to use the TURN server. This defaults to true, otherwise
VoIP will be unreliable for guests. However, it does introduce a slight security risk as
it allows users to connect to arbitrary endpoints without having first signed up for a valid account (e.g. by passing a CAPTCHA).
Enable registration without email or captcha verification. Note: this option is not recommended,
as registration without verification is a known vector for spam and abuse. Defaults to false. Has no effect
unless enable_registration is also enabled.
Mandate that users are only allowed to associate certain formats of
3PIDs with accounts on this server, as specified by the medium and pattern sub-options.
Require users to submit a token during registration.
Tokens can be managed using the admin API.
Disabling this option will not delete any tokens previously generated.
Defaults to false. Set to true to enable.
Note that enable_registration must also be set to allow account registration.
Set the number of bcrypt rounds used to generate password hash.
Larger numbers increase the work factor needed to generate the hash.
The default number is 12 (which equates to 2^12 rounds).
N.B. that increasing this will exponentially increase the time required
to register or login - e.g. 24 => 2^24 rounds which will take >20 mins.
Example configuration:
Allows users to register as guests without a password/email/etc, and
participate in rooms hosted on this server which have been made
accessible to anonymous users. Defaults to false.
Delegate verification of phone numbers to an identity server.
When a user wishes to add a phone number to their account, we need to verify that they
actually own that phone number, which requires sending them a text message (SMS).
Currently Synapse does not support sending those texts itself and instead delegates the
task to an identity server. The base URI for the identity server to be used is
specified by the account_threepid_delegates.msisdn option.
If this is left unspecified, Synapse will not allow users to add phone numbers to
their account.
(Servers handling the these requests must answer the /requestToken endpoints defined
by the Matrix Identity Service API
specification.)
Deprecated in Synapse 1.64.0: The email option is deprecated.
Removed in Synapse 1.66.0: The email option has been removed.
If present, Synapse will report a configuration error on startup.
Example configuration:
account_threepid_delegates:
msisdn: http://localhost:8090 # Delegate SMS sending to this local process
Whether users are allowed to change their displayname after it has
been initially set. Useful when provisioning users based on the
contents of a third-party directory.
Does not apply to server administrators. Defaults to true.
Whether users are allowed to change their avatar after it has been
initially set. Useful when provisioning users based on the contents
of a third-party directory.
Does not apply to server administrators. Defaults to true.
Users who register on this homeserver will automatically be joined
to the rooms listed under this option.
By default, any room aliases included in this list will be created
as a publicly joinable room when the first user registers for the
homeserver. If the room already exists, make certain it is a publicly joinable
room, i.e. the join rule of the room must be set to 'public'. You can find more options
relating to auto-joining rooms below.
As Spaces are just rooms under the hood, Space aliases may also be
used.
Where auto_join_rooms are specified, setting this flag ensures that
the rooms exist by creating them when the first user on the
homeserver registers. This option will not create Spaces.
By default the auto-created rooms are publicly joinable from any federated
server. Use the autocreate_auto_join_rooms_federated and
autocreate_auto_join_room_preset settings to customise this behaviour.
Setting to false means that if the rooms are not manually created,
users cannot be auto-joined since they do not exist.
Whether the rooms listed in auto_join_rooms that are auto-created are available
via federation. Only has an effect if autocreate_auto_join_rooms is true.
Note that whether a room is federated cannot be modified after
creation.
Defaults to true: the room will be joinable from other servers.
Set to false to prevent users from other homeservers from
joining these rooms.
The local part of the user id which is used to create auto_join_rooms if
autocreate_auto_join_rooms is true. If this is not provided then the
initial user account that registers will be used to create the rooms.
The user id is also used to invite new users to any auto-join rooms which
are set to invite-only.
It must be configured if autocreate_auto_join_room_preset is set to
"private_chat" or "trusted_private_chat".
Note that this must be specified in order for new users to be correctly
invited to any auto-join rooms which have been set to invite-only (either
at the time of creation or subsequently).
Note that, if the room already exists, this user must be joined and
have the appropriate permissions to invite new members.
Whether to inhibit errors raised when registering a new account if the user ID
already exists. If turned on, requests to /register/available will always
show a user ID as available, and Synapse won't raise an error when starting
a registration with a user ID that already exists. However, Synapse will still
raise an error if the registration completes and the username conflicts.
Time that a refresh token remains valid for (provided that it is not
exchanged for another one first).
This option can be used to automatically log-out inactive sessions.
Please see the manual for more information.
Note also that this is calculated at login time and refresh time:
changes are not applied to existing sessions until they are refreshed.
Time that an access token remains valid for, if the session is NOT
using refresh tokens.
Please note that not all clients support refresh tokens, so setting
this to a short value may be inconvenient for some users who will
then be logged out frequently.
Note also that this is calculated at login time: changes are not applied
retrospectively to existing sessions for users that have already logged in.
The amount of time to allow a user-interactive authentication session to be active.
This defaults to 0, meaning the user is queried for their credentials
before every action, but this can be overridden to allow a single
validation to be re-used. This weakens the protections afforded by
the user-interactive authentication process, by allowing for multiple
(and potentially different) operations to use the same validation session.
This is ignored for potentially "dangerous" operations (including
deactivating an account, modifying an account password, adding a 3PID,
and minting additional login tokens).
Use the session_timeout sub-option here to change the time allowed for credential validation.
Use this option to enable sentry integration. Provide the DSN assigned to you by sentry
with the dsn setting.
NOTE: While attempts are made to ensure that the logs don't contain
any sensitive information, this cannot be guaranteed. By enabling
this option the sentry server may therefore receive sensitive
information, and it in turn may then disseminate sensitive information
through insecure notification channels if so configured.
Flags to enable Prometheus metrics which are not suitable to be
enabled by default, either for performance reasons or limited use.
Currently the only option is known_servers, which publishes
synapse_federation_known_servers, a gauge of the number of
servers this homeserver knows about, including itself. May cause
performance problems on large homeservers.
Whether or not to report homeserver usage statistics. This is originally
set when generating the config. Set this option to true or false to change the current
behavior. See
Reporting Homeserver Usage Statistics
for information on what data is reported.
Statistics will be reported 5 minutes after Synapse starts, and then every 3 hours
after that.
This setting controls the state that is shared with users upon receiving an
invite to a room, or in reply to a knock on a room. By default, the following
state events are shared with users:
m.room.join_rules
m.room.canonical_alias
m.room.avatar
m.room.encryption
m.room.name
m.room.create
m.room.topic
To change the default behavior, use the following sub-options:
disable_default_event_types: boolean. Set to true to disable the above
defaults. If this is enabled, only the event types listed in
additional_event_types are shared. Defaults to false.
additional_event_types: A list of additional state events to include in the
events to be shared. By default, this list is empty (so only the default event
types are shared).
Each entry in this list should be either a single string or a list of two
strings.
A standalone string t represents all events with type t (i.e.
with no restrictions on state keys).
A pair of strings [t, s] represents a single event with type t and
state key s. The same type can appear in two entries with different state
keys: in this situation, both state keys are included in prejoin state.
Example configuration:
room_prejoin_state:
disable_default_event_types: false
additional_event_types:
# Share all events of type `org.example.custom.event.typeA`
- org.example.custom.event.typeA
# Share only events of type `org.example.custom.event.typeB` whose
# state_key is "foo"
- ["org.example.custom.event.typeB", "foo"]
# Share only events of type `org.example.custom.event.typeC` whose
# state_key is "bar" or "baz"
- ["org.example.custom.event.typeC", "bar"]
- ["org.example.custom.event.typeC", "baz"]
Changed in Synapse 1.74: admins can filter the events in prejoin state based
on their state key.
We record the IP address of clients used to access the API for various
reasons, including displaying it to the user in the "Where you're signed in"
dialog.
By default, when puppeting another user via the admin API, the client IP
address is recorded against the user who created the access token (ie, the
admin user), and not the puppeted user.
Set this option to true to also record the IP address against the puppeted
user. (This also means that the puppeted user will count as an "active" user
for the purpose of monthly active user tracking - see limit_usage_by_mau etc
above.)
Whether to send the application service access tokens via the access_token query parameter
per older versions of the Matrix specification. Defaults to false. Set to true to enable sending
access tokens via a query parameter.
**Enabling this option is considered insecure and is not recommended. **
The keys that the server used to sign messages with but won't use
to sign new messages. For each key, key should be the base64-encoded public key, and
expired_tsshould be the time (in milliseconds since the unix epoch) that
it was last used.
It is possible to build an entry from an old signing.key file using the
export_signing_key script which is provided with synapse.
How long key response published by this server is valid for.
Used to set the valid_until_ts in /key/v2 APIs.
Determines how quickly servers will query to check which keys
are still valid. Defaults to 1d.
The trusted servers to download signing keys from.
When we need to fetch a signing key, each server is tried in parallel.
Normally, the connection to the key server is validated via TLS certificates.
Additional security can be provided by configuring a verify key, which
will make synapse check that the response is signed by that key.
This setting supersedes an older setting named perspectives. The old format
is still supported for backwards-compatibility, but it is deprecated.
trusted_key_servers defaults to matrix.org, but using it will generate a
warning on start-up. To suppress this warning, set
suppress_key_server_warning to true.
If the use of a trusted key server has to be deactivated, e.g. in a private
federation or for privacy reasons, this can be realised by setting
an empty array (trusted_key_servers: []). Then Synapse will request the keys
directly from the server that owns the keys. If Synapse does not get keys directly
from the server, the events of this server will be rejected.
Options for each entry in the list include:
server_name: the name of the server. Required.
verify_keys: an optional map from key id to base64-encoded public key.
If specified, we will check that the response is signed by at least
one of the given keys.
accept_keys_insecurely: a boolean. Normally, if verify_keys is unset,
and federation_verify_certificates is not true, synapse will refuse
to start, because this would allow anyone who can spoof DNS responses
to masquerade as the trusted key server. If you know what you are doing
and are sure that your network environment provides a secure connection
to the key server, you can set this to true to override this behaviour.
Enable SAML2 for registration and login. Uses pysaml2. To learn more about pysaml and
to find a full list options for configuring pysaml, read the docs here.
At least one of sp_config or config_path must be set in this section to
enable SAML login. You can either put your entire pysaml config inline using the sp_config
option, or you can specify a path to a psyaml config file with the sub-option config_path.
This setting has the following sub-options:
idp_name: A user-facing name for this identity provider, which is used to
offer the user a choice of login mechanisms.
idp_icon: An optional icon for this identity provider, which is presented
by clients and Synapse's own IdP picker page. If given, must be an
MXC URI of the format mxc://<server-name>/<media-id>. (An easy way to
obtain such an MXC URI is to upload an image to an (unencrypted) room
and then copy the "url" from the source of the event.)
idp_brand: An optional brand for this identity provider, allowing clients
to style the login flow according to the identity provider in question.
See the spec for possible options here.
sp_config: the configuration for the pysaml2 Service Provider. See pysaml2 docs for format of config.
Default values will be used for the entityid and service settings,
so it is not normally necessary to specify them unless you need to
override them. Here are a few useful sub-options for configuring pysaml:
metadata: Point this to the IdP's metadata. You must provide either a local
file via the local attribute or (preferably) a URL via the
remote attribute.
accepted_time_diff: 3: Allowed clock difference in seconds between the homeserver and IdP.
Defaults to 0.
service: By default, the user has to go to our login page first. If you'd like
to allow IdP-initiated login, set allow_unsolicited to true under sp in the service
section.
config_path: specify a separate pysaml2 configuration file thusly:
config_path: "CONFDIR/sp_conf.py"
saml_session_lifetime: The lifetime of a SAML session. This defines how long a user has to
complete the authentication process, if allow_unsolicited is unset. The default is 15 minutes.
user_mapping_provider: Using this option, an external module can be provided as a
custom solution to mapping attributes returned from a saml provider onto a matrix user. The
user_mapping_provider has the following attributes:
module: The custom module's class.
config: Custom configuration values for the module. Use the values provided in the
example if you are using the built-in user_mapping_provider, or provide your own
config values for a custom class if you are using one. This section will be passed as a Python
dictionary to the module's parse_config method. The built-in provider takes the following two
options:
mxid_source_attribute: The SAML attribute (after mapping via the attribute maps) to use
to derive the Matrix ID from. It is 'uid' by default. Note: This used to be configured by the
saml2_config.mxid_source_attribute option. If that is still defined, its value will be used instead.
mxid_mapping: The mapping system to use for mapping the saml attribute onto a
matrix ID. Options include: hexencode (which maps unpermitted characters to '=xx')
and dotreplace (which replaces unpermitted characters with '.').
The default is hexencode. Note: This used to be configured by the
saml2_config.mxid_mapping option. If that is still defined, its value will be used instead.
grandfathered_mxid_source_attribute: In previous versions of synapse, the mapping from SAML attribute to
MXID was always calculated dynamically rather than stored in a table. For backwards- compatibility, we will look for user_ids
matching such a pattern before creating a new account. This setting controls the SAML attribute which will be used for this
backwards-compatibility lookup. Typically it should be 'uid', but if the attribute maps are changed, it may be necessary to change it.
The default is 'uid'.
attribute_requirements: It is possible to configure Synapse to only allow logins if SAML attributes
match particular values. The requirements can be listed under
attribute_requirements as shown in the example. All of the listed attributes must
match for the login to be permitted.
idp_entityid: If the metadata XML contains multiple IdP entities then the idp_entityid
option must be set to the entity to redirect users to.
Most deployments only have a single IdP entity and so should omit this option.
Once SAML support is enabled, a metadata file will be exposed at
https://<server>:<port>/_synapse/client/saml2/metadata.xml, which you may be able to
use to configure your SAML IdP with. Alternatively, you can manually configure
the IdP to use an ACS location of
https://<server>:<port>/_synapse/client/saml2/authn_response.
Example configuration:
saml2_config:
sp_config:
metadata:
local: ["saml2/idp.xml"]
remote:
- url: https://our_idp/metadata.xml
accepted_time_diff: 3
service:
sp:
allow_unsolicited: true
# The examples below are just used to generate our metadata xml, and you
# may well not need them, depending on your setup. Alternatively you
# may need a whole lot more detail - see the pysaml2 docs!
description: ["My awesome SP", "en"]
name: ["Test SP", "en"]
ui_info:
display_name:
- lang: en
text: "Display Name is the descriptive name of your service."
description:
- lang: en
text: "Description should be a short paragraph explaining the purpose of the service."
information_url:
- lang: en
text: "https://example.com/terms-of-service"
privacy_statement_url:
- lang: en
text: "https://example.com/privacy-policy"
keywords:
- lang: en
text: ["Matrix", "Element"]
logo:
- lang: en
text: "https://example.com/logo.svg"
width: "200"
height: "80"
organization:
name: Example com
display_name:
- ["Example co", "en"]
url: "http://example.com"
contact_person:
- given_name: Bob
sur_name: "the Sysadmin"
email_address": ["admin@example.com"]
contact_type": technical
saml_session_lifetime: 5m
user_mapping_provider:
# Below options are intended for the built-in provider, they should be
# changed if using a custom module.
config:
mxid_source_attribute: displayName
mxid_mapping: dotreplace
grandfathered_mxid_source_attribute: upn
attribute_requirements:
- attribute: userGroup
value: "staff"
- attribute: department
value: "sales"
idp_entityid: 'https://our_idp/entityid'
List of OpenID Connect (OIDC) / OAuth 2.0 identity providers, for registration
and login. See here
for information on how to configure these options.
For backwards compatibility, it is also possible to configure a single OIDC
provider via an oidc_config setting. This is now deprecated and admins are
advised to migrate to the oidc_providers format. (When doing that migration,
use oidc for the idp_id to ensure that existing users continue to be
recognised.)
Options for each entry include:
idp_id: a unique identifier for this identity provider. Used internally
by Synapse; should be a single word such as 'github'.
Note that, if this is changed, users authenticating via that provider
will no longer be recognised as the same user!
(Use "oidc" here if you are migrating from an old oidc_config configuration.)
idp_name: A user-facing name for this identity provider, which is used to
offer the user a choice of login mechanisms.
idp_icon: An optional icon for this identity provider, which is presented
by clients and Synapse's own IdP picker page. If given, must be an
MXC URI of the format mxc://<server-name>/<media-id>. (An easy way to
obtain such an MXC URI is to upload an image to an (unencrypted) room
and then copy the "url" from the source of the event.)
idp_brand: An optional brand for this identity provider, allowing clients
to style the login flow according to the identity provider in question.
See the spec for possible options here.
discover: set to false to disable the use of the OIDC discovery mechanism
to discover endpoints. Defaults to true.
issuer: Required. The OIDC issuer. Used to validate tokens and (if discovery
is enabled) to discover the provider's endpoints.
client_id: Required. oauth2 client id to use.
client_secret: oauth2 client secret to use. May be omitted if
client_secret_jwt_key is given, or if client_auth_method is 'none'.
Must be omitted if client_secret_path is specified.
client_secret_path: path to the oauth2 client secret to use. With that
it's not necessary to leak secrets into the config file itself.
Mutually exclusive with client_secret. Can be omitted if
client_secret_jwt_key is specified.
Added in Synapse 1.91.0.
client_secret_jwt_key: Alternative to client_secret: details of a key used
to create a JSON Web Token to be used as an OAuth2 client secret. If
given, must be a dictionary with the following properties:
key: a pem-encoded signing key. Must be a suitable key for the
algorithm specified. Required unless key_file is given.
key_file: the path to file containing a pem-encoded signing key file.
Required unless key is given.
jwt_header: a dictionary giving properties to include in the JWT
header. Must include the key alg, giving the algorithm used to
sign the JWT, such as "ES256", using the JWA identifiers in
RFC7518.
jwt_payload: an optional dictionary giving properties to include in
the JWT payload. Normally this should include an iss key.
client_auth_method: auth method to use when exchanging the token. Valid
values are client_secret_basic (default), client_secret_post and
none.
pkce_method: Whether to use proof key for code exchange when requesting
and exchanging the token. Valid values are: auto, always, or never. Defaults
to auto, which uses PKCE if supported during metadata discovery. Set to always
to force enable PKCE or never to force disable PKCE.
scopes: list of scopes to request. This should normally include the "openid"
scope. Defaults to ["openid"].
authorization_endpoint: the oauth2 authorization endpoint. Required if
provider discovery is disabled.
token_endpoint: the oauth2 token endpoint. Required if provider discovery is
disabled.
userinfo_endpoint: the OIDC userinfo endpoint. Required if discovery is
disabled and the 'openid' scope is not requested.
jwks_uri: URI where to fetch the JWKS. Required if discovery is disabled and
the 'openid' scope is used.
skip_verification: set to 'true' to skip metadata verification. Use this if
you are connecting to a provider that is not OpenID Connect compliant.
Defaults to false. Avoid this in production.
user_profile_method: Whether to fetch the user profile from the userinfo
endpoint, or to rely on the data returned in the id_token from the token_endpoint.
Valid values are: auto or userinfo_endpoint.
Defaults to auto, which uses the userinfo endpoint if openid is
not included in scopes. Set to userinfo_endpoint to always use the
userinfo endpoint.
allow_existing_users: set to true to allow a user logging in via OIDC to
match a pre-existing account instead of failing. This could be used if
switching from password logins to OIDC. Defaults to false.
enable_registration: set to 'false' to disable automatic registration of new
users. This allows the OIDC SSO flow to be limited to sign in only, rather than
automatically registering users that have a valid SSO login but do not have
a pre-registered account. Defaults to true.
user_mapping_provider: Configuration for how attributes returned from a OIDC
provider are mapped onto a matrix user. This setting has the following
sub-properties:
module: The class name of a custom mapping module. Default is
synapse.handlers.oidc.JinjaOidcMappingProvider.
See OpenID Mapping Providers
for information on implementing a custom mapping provider.
config: Configuration for the mapping provider module. This section will
be passed as a Python dictionary to the user mapping provider
module's parse_config method.
For the default provider, the following settings are available:
subject_template: Jinja2 template for a unique identifier for the user.
Defaults to {{ user.sub }}, which OpenID Connect compliant providers should provide.
This replaces and overrides subject_claim.
subject_claim: name of the claim containing a unique identifier
for the user. Defaults to 'sub', which OpenID Connect
compliant providers should provide.
Deprecated in Synapse v1.75.0.
picture_template: Jinja2 template for an url for the user's profile picture.
Defaults to {{ user.picture }}, which OpenID Connect compliant providers should
provide and has to refer to a direct image file such as PNG, JPEG, or GIF image file.
This replaces and overrides picture_claim.
Currently only supported in monolithic (single-process) server configurations
where the media repository runs within the Synapse process.
picture_claim: name of the claim containing an url for the user's profile picture.
Defaults to 'picture', which OpenID Connect compliant providers should provide
and has to refer to a direct image file such as PNG, JPEG, or GIF image file.
Currently only supported in monolithic (single-process) server configurations
where the media repository runs within the Synapse process.
Deprecated in Synapse v1.75.0.
localpart_template: Jinja2 template for the localpart of the MXID.
If this is not set, the user will be prompted to choose their
own username (see the documentation for the sso_auth_account_details.html
template). This template can use the localpart_from_email filter.
confirm_localpart: Whether to prompt the user to validate (or
change) the generated localpart (see the documentation for the
'sso_auth_account_details.html' template), instead of
registering the account right away.
display_name_template: Jinja2 template for the display name to set
on first login. If unset, no displayname will be set.
email_template: Jinja2 template for the email address of the user.
If unset, no email address will be added to the account.
extra_attributes: a map of Jinja2 templates for extra attributes
to send back to the client during login. Note that these are non-standard and clients will ignore them
without modifications.
When rendering, the Jinja2 templates are given a 'user' variable,
which is set to the claims returned by the UserInfo Endpoint and/or
in the ID Token.
backchannel_logout_enabled: set to true to process OIDC Back-Channel Logout notifications.
Those notifications are expected to be received on /_synapse/client/oidc/backchannel_logout.
Defaults to false.
backchannel_logout_ignore_sub: by default, the OIDC Back-Channel Logout feature checks that the
sub claim matches the subject claim received during login. This check can be disabled by setting
this to true. Defaults to false.
You might want to disable this if the subject_claim returned by the mapping provider is not sub.
It is possible to configure Synapse to only allow logins if certain attributes
match particular values in the OIDC userinfo. The requirements can be listed under
attribute_requirements as shown here:
All of the listed attributes must match for the login to be permitted. Additional attributes can be added to
userinfo by expanding the scopes section of the OIDC config to retrieve
additional information from the OIDC provider.
If the OIDC claim is a list, then the attribute must match any value in the list.
Otherwise, it must exactly match the value of the claim. Using the example
above, the family_name claim MUST be "Stephensson", but the groups
claim MUST contain "admin".
Enable Central Authentication Service (CAS) for registration and login.
Has the following sub-options:
enabled: Set this to true to enable authorization against a CAS server.
Defaults to false.
idp_name: A user-facing name for this identity provider, which is used to
offer the user a choice of login mechanisms.
idp_icon: An optional icon for this identity provider, which is presented
by clients and Synapse's own IdP picker page. If given, must be an
MXC URI of the format mxc://<server-name>/<media-id>. (An easy way to
obtain such an MXC URI is to upload an image to an (unencrypted) room
and then copy the "url" from the source of the event.)
idp_brand: An optional brand for this identity provider, allowing clients
to style the login flow according to the identity provider in question.
See the spec for possible options here.
server_url: The URL of the CAS authorization endpoint.
protocol_version: The CAS protocol version, defaults to none (version 3 is required if you want to use "required_attributes").
displayname_attribute: The attribute of the CAS response to use as the display name.
If no name is given here, no displayname will be set.
required_attributes: It is possible to configure Synapse to only allow logins if CAS attributes
match particular values. All of the keys given below must exist
and the values must match the given value. Alternately if the given value
is None then any value is allowed (the attribute just must exist).
All of the listed attributes must match for the login to be permitted.
enable_registration: set to 'false' to disable automatic registration of new
users. This allows the CAS SSO flow to be limited to sign in only, rather than
automatically registering users that have a valid SSO login but do not have
a pre-registered account. Defaults to true.
Additional settings to use with single-sign on systems such as OpenID Connect,
SAML2 and CAS.
Server admins can configure custom templates for pages related to SSO. See
here for more information.
Options include:
client_whitelist: A list of client URLs which are whitelisted so that the user does not
have to confirm giving access to their account to the URL. Any client
whose URL starts with an entry in the following list will not be subject
to an additional confirmation step after the SSO login is completed.
WARNING: An entry such as "https://my.client" is insecure, because it
will also match "https://my.client.evil.site", exposing your users to
phishing attacks from evil.site. To avoid this, include a slash after the
hostname: "https://my.client/".
The login fallback page (used by clients that don't natively support the
required login flows) is whitelisted in addition to any URLs in this list.
By default, this list contains only the login fallback page.
update_profile_information: Use this setting to keep a user's profile fields in sync with information from
the identity provider. Currently only syncing the displayname is supported. Fields
are checked on every SSO login, and are updated if necessary.
Note that enabling this option will override user profile information,
regardless of whether users have opted-out of syncing that
information when first signing in. Defaults to false.
JSON web token integration. The following settings can be used to make
Synapse JSON web tokens for authentication, instead of its internal
password database.
Each JSON Web Token needs to contain a "sub" (subject) claim, which is
used as the localpart of the mxid.
Additionally, the expiration time ("exp"), not before time ("nbf"),
and issued at ("iat") claims are validated if present.
Note that this is a non-standard login type and client support is
expected to be non-existent.
enabled: Set to true to enable authorization using JSON web
tokens. Defaults to false.
secret: This is either the private shared secret or the public key used to
decode the contents of the JSON web token. Required if enabled is set to true.
algorithm: The algorithm used to sign (or HMAC) the JSON web token.
Supported algorithms are listed
here (section JWS).
Required if enabled is set to true.
subject_claim: Name of the claim containing a unique identifier for the user.
Optional, defaults to sub.
issuer: The issuer to validate the "iss" claim against. Optional. If provided the
"iss" claim will be required and validated for all JSON web tokens.
audiences: A list of audiences to validate the "aud" claim against. Optional.
If provided the "aud" claim will be required and validated for all JSON web tokens.
Note that if the "aud" claim is included in a JSON web token then
validation will fail without configuring audiences.
enabled: Defaults to true.
Set to false to disable password authentication.
Set to only_for_reauth to allow users with existing passwords to use them
to log in and reauthenticate, whilst preventing new users from setting passwords.
localdb_enabled: Set to false to disable authentication against the local password
database. This is ignored if enabled is false, and is only useful
if you have other password_providers. Defaults to true.
pepper: Set the value here to a secret random string for extra security.
DO NOT CHANGE THIS AFTER INITIAL SETUP!
policy: Define and enforce a password policy, such as minimum lengths for passwords, etc.
Each parameter is optional. This is an implementation of MSC2000. Parameters are as follows:
enabled: Defaults to false. Set to true to enable.
minimum_length: Minimum accepted length for a password. Defaults to 0.
require_digit: Whether a password must contain at least one digit.
Defaults to false.
require_symbol: Whether a password must contain at least one symbol.
A symbol is any character that's not a number or a letter. Defaults to false.
require_lowercase: Whether a password must contain at least one lowercase letter.
Defaults to false.
require_uppercase: Whether a password must contain at least one uppercase letter.
Defaults to false.
This setting defines options for push notifications.
This option has a number of sub-options. They are as follows:
enabled: Enables or disables push notification calculation. Note, disabling this will also
stop unread counts being calculated for rooms. This mode of operation is intended
for homeservers which may only have bots or appservice users connected, or are otherwise
not interested in push/unread counters. This is enabled by default.
include_content: Clients requesting push notifications can either have the body of
the message sent in the notification poke along with other details
like the sender, or just the event ID and room ID (event_id_only).
If clients choose the to have the body sent, this option controls whether the
notification request includes the content of the event (other details
like the sender are still included). If event_id_only is enabled, it
has no effect.
For modern android devices the notification content will still appear
because it is loaded by the app. iPhone, however will send a
notification saying only that a message arrived and who it came from.
Defaults to true. Set to false to only include the event ID and room ID in push notification payloads.
group_unread_count_by_room: false: When a push notification is received, an unread count is also sent.
This number can either be calculated as the number of unread messages for the user, or the number of rooms the
user has unread messages in. Defaults to true, meaning push clients will see the number of
rooms with unread messages in them. Set to false to instead send the number
of unread messages.
jitter_delay: Delays push notifications by a random amount up to the given
duration. Useful for mitigating timing attacks. Optional, defaults to no
delay. Added in Synapse 1.84.0.
This setting defines options related to the user directory.
This option has the following sub-options:
enabled: Defines whether users can search the user directory. If false then
empty responses are returned to all queries. Defaults to true.
search_all_users: Defines whether to search all users visible to your HS at the time the search is performed. If set to true, will return all users who share a room with the user from the homeserver.
If false, search results will only contain users
visible in public rooms and users sharing a room with the requester.
Defaults to false.
NB. If you set this to true, and the last time the user_directory search
indexes were (re)built was before Synapse 1.44, you'll have to
rebuild the indexes in order to search through all known users.
These indexes are built the first time Synapse starts; admins can
manually trigger a rebuild via the API following the instructions
for running background updates,
set to true to return search results containing all known users, even if that
user does not share a room with the requester.
prefer_local_users: Defines whether to prefer local users in search query results.
If set to true, local users are more likely to appear above remote users when searching the
user directory. Defaults to false.
show_locked_users: Defines whether to show locked users in search query results. Defaults to false.
For detailed instructions on user consent configuration, see here.
Parts of this section are required if enabling the consent resource under
listeners, in particular template_dir and version.
template_dir: gives the location of the templates for the HTML forms.
This directory should contain one subdirectory per language (eg, en, fr),
and each language directory should contain the policy document (named as
.html) and a success page (success.html).
version: specifies the 'current' version of the policy document. It defines
the version to be served by the consent resource if there is no 'v'
parameter.
server_notice_content: if enabled, will send a user a "Server Notice"
asking them to consent to the privacy policy. The server_notices section
must also be configured for this to work. Notices will not be sent to
guest users unless send_server_notice_to_guests is set to true.
block_events_error, if set, will block any attempts to send events
until the user consents to the privacy policy. The value of the setting is
used as the text of the error.
require_at_registration, if enabled, will add a step to the registration
process, similar to how captcha works. Users will be required to accept the
policy before their account is created.
policy_name is the display name of the policy users will see when registering
for an account. Has no effect unless require_at_registration is enabled.
Defaults to "Privacy Policy".
Example configuration:
user_consent:
template_dir: res/templates/privacy
version: 1.0
server_notice_content:
msgtype: m.text
body: >-
To continue using this homeserver you must review and agree to the
terms and conditions at %(consent_uri)s
send_server_notice_to_guests: true
block_events_error: >-
To continue using this homeserver you must review and agree to the
terms and conditions at %(consent_uri)s
require_at_registration: false
policy_name: Privacy Policy
Settings for local room and user statistics collection. See here
for more.
enabled: Set to false to disable room and user statistics. Note that doing
so may cause certain features (such as the room directory) not to work
correctly. Defaults to true.
Use this setting to enable a room which can be used to send notices
from the server to users. It is a special room which users cannot leave; notices
in the room come from a special "notices" user id.
If you use this setting, you must define the system_mxid_localpart
sub-setting, which defines the id of the user which will be used to send the
notices.
Sub-options for this setting include:
system_mxid_display_name: set the display name of the "notices" user
system_mxid_avatar_url: set the avatar for the "notices" user
room_name: set the room name of the server notices room
auto_join: boolean. If true, the user will be automatically joined to the room instead of being invited.
Defaults to false. Added in Synapse 1.98.0.
Set to false to disable searching the public room list. When disabled
blocks searching local and remote room lists for local and remote
users by always returning an empty list for all queries. Defaults to true.
The alias_creation_rules option allows server admins to prevent unwanted
alias creation on this server.
This setting is an optional list of 0 or more rules. By default, no list is
provided, meaning that all alias creations are permitted.
Otherwise, requests to create aliases are matched against each rule in order.
The first rule that matches decides if the request is allowed or denied. If no
rule matches, the request is denied. In particular, this means that configuring
an empty list of rules will deny every alias creation request.
Each rule is a YAML object containing four fields, each of which is an optional string:
user_id: a glob pattern that matches against the creator of the alias.
alias: a glob pattern that matches against the alias being created.
room_id: a glob pattern that matches against the room ID the alias is being pointed at.
action: either allow or deny. What to do with the request if the rule matches. Defaults to allow.
Each of the glob patterns is optional, defaulting to * ("match anything").
Note that the patterns match against fully qualified IDs, e.g. against
@alice:example.com, #room:example.com and !abcdefghijk:example.com instead
of alice, room and abcedgghijk.
Example configuration:
# No rule list specified. All alias creations are allowed.
# This is the default behaviour.
alias_creation_rules:
# A list of one rule which allows everything.
# This has the same effect as the previous example.
alias_creation_rules:
- "action": "allow"
# An empty list of rules. All alias creations are denied.
alias_creation_rules: []
# A list of one rule which denies everything.
# This has the same effect as the previous example.
alias_creation_rules:
- "action": "deny"
# Prevent a specific user from creating aliases.
# Allow other users to create any alias
alias_creation_rules:
- user_id: "@bad_user:example.com"
action: deny
- action: allow
# Prevent aliases being created which point to a specific room.
alias_creation_rules:
- room_id: "!forbiddenRoom:example.com"
action: deny
- action: allow
The room_list_publication_rules option allows server admins to prevent
unwanted entries from being published in the public room list.
The format of this option is the same as that for
alias_creation_rules: an optional list of 0 or more
rules. By default, no list is provided, meaning that all rooms may be
published to the room list.
Otherwise, requests to publish a room are matched against each rule in order.
The first rule that matches decides if the request is allowed or denied. If no
rule matches, the request is denied. In particular, this means that configuring
an empty list of rules will deny every alias creation request.
Each rule is a YAML object containing four fields, each of which is an optional string:
user_id: a glob pattern that matches against the user publishing the room.
alias: a glob pattern that matches against one of published room's aliases.
If the room has no aliases, the alias match fails unless alias is unspecified or *.
If the room has exactly one alias, the alias match succeeds if the alias pattern matches that alias.
If the room has two or more aliases, the alias match succeeds if the pattern matches at least one of the aliases.
room_id: a glob pattern that matches against the room ID of the room being published.
action: either allow or deny. What to do with the request if the rule matches. Defaults to allow.
Each of the glob patterns is optional, defaulting to * ("match anything").
Note that the patterns match against fully qualified IDs, e.g. against
@alice:example.com, #room:example.com and !abcdefghijk:example.com instead
of alice, room and abcedgghijk.
Example configuration:
# No rule list specified. Anyone may publish any room to the public list.
# This is the default behaviour.
room_list_publication_rules:
# A list of one rule which allows everything.
# This has the same effect as the previous example.
room_list_publication_rules:
- "action": "allow"
# An empty list of rules. No-one may publish to the room list.
room_list_publication_rules: []
# A list of one rule which denies everything.
# This has the same effect as the previous example.
room_list_publication_rules:
- "action": "deny"
# Prevent a specific user from publishing rooms.
# Allow other users to publish anything.
room_list_publication_rules:
- user_id: "@bad_user:example.com"
action: deny
- action: allow
# Prevent publication of a specific room.
room_list_publication_rules:
- room_id: "!forbiddenRoom:example.com"
action: deny
- action: allow
# Prevent publication of rooms with at least one alias containing the word "potato".
room_list_publication_rules:
- alias: "#*potato*:example.com"
action: deny
- action: allow
The default_power_level_content_override option controls the default power
levels for rooms.
Useful if you know that your users need special permissions in rooms
that they create (e.g. to send particular types of state events without
needing an elevated power level). This takes the same shape as the
power_level_content_override parameter in the /createRoom API, but
is applied before that parameter.
Note that each key provided inside a preset (for example events in the example
below) will overwrite all existing defaults inside that key. So in the example
below, newly-created private_chat rooms will have no rules for any event types
except com.example.foo.
A list of rooms to exclude from sync responses. This is useful for server
administrators wishing to group users into a room without these users being able
to see it from their client.
These settings enable and configure opentracing, which implements distributed tracing.
This allows you to observe the causal chains of events across servers
including requests, key lookups etc., across any server running
synapse or any other services which support opentracing
(specifically those implemented with Jaeger).
Sub-options include:
enabled: whether tracing is enabled. Set to true to enable. Disabled by default.
homeserver_whitelist: The list of homeservers we wish to send and receive span contexts and span baggage.
See here for more.
This is a list of regexes which are matched against the server_name of the homeserver.
By default, it is empty, so no servers are matched.
force_tracing_for_users: # A list of the matrix IDs of users whose requests will always be traced,
even if the tracing system would otherwise drop the traces due to probabilistic sampling.
By default, the list is empty.
jaeger_config: Jaeger can be configured to sample traces at different rates.
All configuration options provided by Jaeger can be set here. Jaeger's configuration is
mostly related to trace sampling which is documented here.
Configuration options related to workers which belong in the main config file
(usually called homeserver.yaml).
A Synapse deployment can scale horizontally by running multiple Synapse processes
called workers. Incoming requests are distributed between workers to handle higher
loads. Some workers are privileged and can accept requests from other workers.
As a result, the worker configuration is divided into two parts.
The first part (in this section of the manual) defines which shardable tasks
are delegated to privileged workers. This allows unprivileged workers to make
requests to a privileged worker to act on their behalf.
The second part
controls the behaviour of individual workers in isolation.
It is possible to scale the processes that handle sending push notifications to sygnal
and email by running a generic_worker and adding it's worker_name to
a pusher_instances map. Doing so will remove handling of this function from the main
process. Multiple workers can be added to this map, in which case the work is balanced
across them. Ensure the main process and all pusher workers are restarted after changing
this option.
It is possible to scale the processes that handle sending outbound federation requests
by running a generic_worker and adding it's worker_name to
a federation_sender_instances map. Doing so will remove handling of this function from
the main process. Multiple workers can be added to this map, in which case the work is
balanced across them.
This configuration setting must be shared between all workers handling federation
sending, and if changed all federation sender workers must be stopped at the same time
and then started, to ensure that all instances are running with the same config (otherwise
events may be dropped).
When using workers this should be a map from worker_name to the HTTP
replication listener of the worker, if configured, and to the main process. Each worker
declared under stream_writers and
outbound_federation_restricted_to needs a HTTP
replication listener, and that listener should be included in the instance_map. The
main process also needs an entry on the instance_map, and it should be listed under
mainif even one other worker exists. Ensure the port matches with what is
declared inside the listener block for a replication listener.
Experimental: When using workers you can define which workers should
handle writing to streams such as event persistence and typing notifications.
Any worker specified here must also be in the instance_map.
When using workers, you can restrict outbound federation traffic to only go through a
specific subset of workers. Any worker specified here must also be in the
instance_map.
worker_replication_secret must also be configured to
authorize inter-worker communication.
The worker that is used to run
background tasks for media repository. If running multiple media repositories
you must configure a single instance to run the background tasks. If not provided
this defaults to the main process or your single media_repository worker.
A unique name for the worker. The worker needs a name to be addressed in
further parameters and identification in log files. We strongly recommend
giving each worker a unique worker_name.
A worker can handle HTTP requests. To do so, a worker_listeners option
must be declared, in the same way as the listeners option
in the shared config.
Workers declared in stream_writers and instance_map
will need to include a replication listener here, in order to accept internal HTTP
requests from other workers.
Specifies whether the worker should be started as a daemon process.
If Synapse is being managed by systemd, this option
must be omitted or set to false.
Background updates are database updates that are run in the background in batches.
The duration, minimum batch size, default batch size, whether to sleep between batches and if so, how long to
sleep can all be configured. This is helpful to speed up or slow down the updates.
This setting has the following sub-options:
background_update_duration_ms: How long in milliseconds to run a batch of background updates for. Defaults to 100.
Set a different time to change the default.
sleep_enabled: Whether to sleep between updates. Defaults to true. Set to false to change the default.
sleep_duration_ms: If sleeping between updates, how long in milliseconds to sleep for. Defaults to 1000.
Set a duration to change the default.
min_batch_size: Minimum size a batch of background updates can be. Must be greater than 0. Defaults to 1.
Set a size to change the default.
default_batch_size: The batch size to use for the first iteration of a new background update. The default is 100.
Set a size to change the default.
Below is a sample homeserver configuration file. The homeserver configuration file
can be tweaked to change the behaviour of your homeserver. A restart of the server is
generally required to apply any changes made to this file.
Note that the contents below are not intended to be copied and used as the basis for
a real homeserver.yaml. Instead, if you are starting from scratch, please generate
a fresh config using Synapse by following the instructions in
Installation.
Documentation for all configuration options can be found in the
Configuration Manual.
# This file is maintained as an up-to-date snapshot of the default
# homeserver.yaml configuration generated by Synapse. You can find a
# complete accounting of possible configuration options at
# https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html
#
# It is *not* intended to be copied and used as the basis for a real
# homeserver.yaml. Instead, if you are starting from scratch, please generate
# a fresh config using Synapse by following the instructions in
# https://matrix-org.github.io/synapse/latest/setup/installation.html.
#
################################################################################
# Configuration file for Synapse.
#
# This is a YAML file: see [1] for a quick introduction. Note in particular
# that *indentation is important*: all the elements of a list or dictionary
# should have the same indentation.
#
# [1] https://docs.ansible.com/ansible/latest/reference_appendices/YAMLSyntax.html
#
# For more information on how to configure Synapse, including a complete accounting of
# each option, go to docs/usage/configuration/config_documentation.md or
# https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html
server_name: "SERVERNAME"
pid_file: DATADIR/homeserver.pid
listeners:
- port: 8008
tls: false
type: http
x_forwarded: true
bind_addresses: ['::1', '127.0.0.1']
resources:
- names: [client, federation]
compress: false
database:
name: sqlite3
args:
database: DATADIR/homeserver.db
log_config: "CONFDIR/SERVERNAME.log.config"
media_store_path: DATADIR/media_store
signing_key_path: "CONFDIR/SERVERNAME.signing.key"
trusted_key_servers:
- server_name: "matrix.org"
Below is a sample logging configuration file. This file can be tweaked to control how your
homeserver will output logs. The value of the log_config option in your homeserver config
should be the path to this file.
To apply changes made to this file, send Synapse a SIGHUP signal (or, if using systemd, run
systemctl reload on the Synapse service).
Note that a default logging configuration (shown below) is created automatically alongside
the homeserver config when following the installation instructions.
It should be named <SERVERNAME>.log.config by default.
# Log configuration for Synapse.
#
# This is a YAML file containing a standard Python logging configuration
# dictionary. See [1] for details on the valid settings.
#
# Synapse also supports structured logging for machine readable logs which can
# be ingested by ELK stacks. See [2] for details.
#
# [1]: https://docs.python.org/3/library/logging.config.html#configuration-dictionary-schema
# [2]: https://matrix-org.github.io/synapse/latest/structured_logging.html
version: 1
formatters:
precise:
format: '%(asctime)s - %(name)s - %(lineno)d - %(levelname)s - %(request)s - %(message)s'
handlers:
file:
class: logging.handlers.TimedRotatingFileHandler
formatter: precise
filename: /var/log/matrix-synapse/homeserver.log
when: midnight
backupCount: 3 # Does not include the current log file.
encoding: utf8
# Default to buffering writes to log file for efficiency.
# WARNING/ERROR logs will still be flushed immediately, but there will be a
# delay (of up to `period` seconds, or until the buffer is full with
# `capacity` messages) before INFO/DEBUG logs get written.
buffer:
class: synapse.logging.handlers.PeriodicallyFlushingMemoryHandler
target: file
# The capacity is the maximum number of log lines that are buffered
# before being written to disk. Increasing this will lead to better
# performance, at the expensive of it taking longer for log lines to
# be written to disk.
# This parameter is required.
capacity: 10
# Logs with a level at or above the flush level will cause the buffer to
# be flushed immediately.
# Default value: 40 (ERROR)
# Other values: 50 (CRITICAL), 30 (WARNING), 20 (INFO), 10 (DEBUG)
flushLevel: 30 # Flush immediately for WARNING logs and higher
# The period of time, in seconds, between forced flushes.
# Messages will not be delayed for longer than this time.
# Default value: 5 seconds
period: 5
# A handler that writes logs to stderr. Unused by default, but can be used
# instead of "buffer" and "file" in the logger handlers.
console:
class: logging.StreamHandler
formatter: precise
loggers:
synapse.storage.SQL:
# beware: increasing this to DEBUG will make synapse log sensitive
# information such as access tokens.
level: INFO
root:
level: INFO
# Write logs to the `buffer` handler, which will buffer them together in memory,
# then write them to a file.
#
# Replace "buffer" with "console" to log to stderr instead.
#
handlers: [buffer]
disable_existing_loggers: false
A structured logging system can be useful when your logs are destined for a
machine to parse and process. By maintaining its machine-readable characteristics,
it enables more efficient searching and aggregations when consumed by software
such as the ELK stack.
Synapse's structured logging system is configured via the file that Synapse's
log_config config option points to. The file should include a formatter which
uses the synapse.logging.TerseJsonFormatter class included with Synapse and a
handler which uses the above formatter.
There is also a synapse.logging.JsonFormatter option which does not include
a timestamp in the resulting JSON. This is useful if the log ingester adds its
own timestamp.
A structured logging configuration looks similar to the following:
version: 1
formatters:
structured:
class: synapse.logging.TerseJsonFormatter
handlers:
file:
class: logging.handlers.TimedRotatingFileHandler
formatter: structured
filename: /path/to/my/logs/homeserver.log
when: midnight
backupCount: 3 # Does not include the current log file.
encoding: utf8
loggers:
synapse:
level: INFO
handlers: [remote]
synapse.storage.SQL:
level: WARNING
The above logging config will set Synapse as 'INFO' logging level by default,
with the SQL layer at 'WARNING', and will log to a file, stored as JSON.
It is also possible to configure Synapse to log to a remote endpoint by using the
synapse.logging.RemoteHandler class included with Synapse. It takes the
following arguments:
host: Hostname or IP address of the log aggregator.
port: Numerical port to contact on the host.
maximum_buffer: (Optional, defaults to 1000) The maximum buffer size to allow.
A remote structured logging configuration looks similar to the following:
The above logging config will set Synapse as 'INFO' logging level by default,
with the SQL layer at 'WARNING', and will log JSON formatted messages to a
remote endpoint at 10.1.2.3:9999.
Synapse uses parametrised templates to generate the content of emails it sends and
webpages it shows to users.
By default, Synapse will use the templates listed here.
Server admins can configure an additional directory for Synapse to look for templates
in, allowing them to specify custom templates:
If this setting is not set, or the files named below are not found within the directory,
default templates from within the Synapse package will be used.
Templates that are given variables when being rendered are rendered using Jinja 2.
Templates rendered by Jinja 2 can also access two functions on top of the functions
already available as part of Jinja 2:
Below are the templates Synapse will look for when generating the content of an email:
notif_mail.html and notif_mail.txt: The contents of email notifications of missed
events.
When rendering, this template is given the following variables:
user_display_name: the display name for the user receiving the notification
unsubscribe_link: the link users can click to unsubscribe from email notifications
summary_text: a summary of the notification(s). The text used can be customised
by configuring the various settings in the email.subjects section of the
configuration file.
rooms: a list of rooms containing events to include in the email. Each element is
an object with the following attributes:
title: a human-readable name for the room
hash: a hash of the ID of the room
invite: a boolean, which is True if the room is an invite the user hasn't
accepted yet, False otherwise
notifs: a list of events, or an empty list if invite is True. Each element
is an object with the following attributes:
link: a matrix.to link to the event
ts: the time in milliseconds at which the event was received
messages: a list of messages containing one message before the event, the
message in the event, and one message after the event. Each element is an
object with the following attributes:
event_type: the type of the event
is_historical: a boolean, which is False if the message is the one
that triggered the notification, True otherwise
id: the ID of the event
ts: the time in milliseconds at which the event was sent
sender_name: the display name for the event's sender
sender_avatar_url: the avatar URL (as a mxc:// URL) for the event's
sender
sender_hash: a hash of the user ID of the sender
msgtype: the type of the message
body_text_html: html representation of the message
body_text_plain: plaintext representation of the message
image_url: mxc url of an image, when "msgtype" is "m.image"
link: a matrix.to link to the room
avator_url: url to the room's avator
reason: information on the event that triggered the email to be sent. It's an
object with the following attributes:
room_id: the ID of the room the event was sent in
room_name: a human-readable name for the room the event was sent in
now: the current time in milliseconds
received_at: the time in milliseconds at which the event was received
delay_before_mail_ms: the amount of time in milliseconds Synapse always waits
before ever emailing about a notification (to give the user a chance to respond
to other push or notice the window)
last_sent_ts: the time in milliseconds at which a notification was last sent
for an event in this room
throttle_ms: the minimum amount of time in milliseconds between two
notifications can be sent for this room
password_reset.html and password_reset.txt: The contents of password reset emails
sent by the homeserver.
When rendering, these templates are given a link variable which contains the link the
user must click in order to reset their password.
registration.html and registration.txt: The contents of address verification emails
sent during registration.
When rendering, these templates are given a link variable which contains the link the
user must click in order to validate their email address.
add_threepid.html and add_threepid.txt: The contents of address verification emails
sent when an address is added to a Matrix account.
When rendering, these templates are given a link variable which contains the link the
user must click in order to validate their email address.
Below are the templates Synapse will look for when generating pages related to
registration and password reset:
password_reset_confirmation.html: An HTML page that a user will see when they follow
the link in the password reset email. The user will be asked to confirm the action
before their password is reset.
When rendering, this template is given the following variables:
sid: the session ID for the password reset
token: the token for the password reset
client_secret: the client secret for the password reset
password_reset_success.html and password_reset_failure.html: HTML pages for success
and failure that a user will see when they confirm the password reset flow using the
page above.
When rendering, password_reset_success.html is given no variable, and
password_reset_failure.html is given a failure_reason, which contains the reason
for the password reset failure.
registration_success.html and registration_failure.html: HTML pages for success and
failure that a user will see when they follow the link in an address verification email
sent during registration.
When rendering, registration_success.html is given no variable, and
registration_failure.html is given a failure_reason, which contains the reason
for the registration failure.
add_threepid_success.html and add_threepid_failure.html: HTML pages for success and
failure that a user will see when they follow the link in an address verification email
sent when an address is added to a Matrix account.
When rendering, add_threepid_success.html is given no variable, and
add_threepid_failure.html is given a failure_reason, which contains the reason
for the registration failure.
Below are the templates Synapse will look for when generating pages related to SSO:
sso_login_idp_picker.html: HTML page to prompt the user to choose an
Identity Provider during login.
This is only used if multiple SSO Identity Providers are configured.
When rendering, this template is given the following variables:
redirect_url: the URL that the user will be redirected to after
login.
server_name: the homeserver's name.
providers: a list of available Identity Providers. Each element is
an object with the following attributes:
idp_id: unique identifier for the IdP
idp_name: user-facing name for the IdP
idp_icon: if specified in the IdP config, an MXC URI for an icon
for the IdP
idp_brand: if specified in the IdP config, a textual identifier
for the brand of the IdP
The rendered HTML page should contain a form which submits its results
back as a GET request, with the following query parameters:
redirectUrl: the client redirect URI (ie, the redirect_url passed
to the template)
idp: the 'idp_id' of the chosen IDP.
sso_auth_account_details.html: HTML page to prompt new users to enter a
userid and confirm other details. This is only shown if the
SSO implementation (with any user_mapping_provider) does not return
a localpart.
When rendering, this template is given the following variables:
server_name: the homeserver's name.
idp: details of the SSO Identity Provider that the user logged in
with: an object with the following attributes:
idp_id: unique identifier for the IdP
idp_name: user-facing name for the IdP
idp_icon: if specified in the IdP config, an MXC URI for an icon
for the IdP
idp_brand: if specified in the IdP config, a textual identifier
for the brand of the IdP
user_attributes: an object containing details about the user that
we received from the IdP. May have the following attributes:
display_name: the user's display name
emails: a list of email addresses
localpart: the local part of the Matrix user ID to register,
if localpart_template is set in the mapping provider configuration (empty
string if not)
The template should render a form which submits the following fields:
username: the localpart of the user's chosen user id
sso_new_user_consent.html: HTML page allowing the user to consent to the
server's terms and conditions. This is only shown for new users, and only if
user_consent.require_at_registration is set.
When rendering, this template is given the following variables:
server_name: the homeserver's name.
user_id: the user's matrix proposed ID.
user_profile.display_name: the user's proposed display name, if any.
consent_version: the version of the terms that the user will be
shown
terms_url: a link to the page showing the terms.
The template should render a form which submits the following fields:
accepted_version: the version of the terms accepted by the user
(ie, 'consent_version' from the input variables).
sso_redirect_confirm.html: HTML page for a confirmation step before redirecting back
to the client with the login token.
When rendering, this template is given the following variables:
redirect_url: the URL the user is about to be redirected to.
display_url: the same as redirect_url, but with the query
parameters stripped. The intention is to have a
human-readable URL to show to users, not to use it as
the final address to redirect to.
server_name: the homeserver's name.
new_user: a boolean indicating whether this is the user's first time
logging in.
user_id: the user's matrix ID.
user_profile.avatar_url: an MXC URI for the user's avatar, if any.
None if the user has not set an avatar.
user_profile.display_name: the user's display name. None if the user
has not set a display name.
sso_auth_confirm.html: HTML page which notifies the user that they are authenticating
to confirm an operation on their account during the user interactive authentication
process.
When rendering, this template is given the following variables:
redirect_url: the URL the user is about to be redirected to.
description: the operation which the user is being asked to confirm
idp: details of the Identity Provider that we will use to confirm
the user's identity: an object with the following attributes:
idp_id: unique identifier for the IdP
idp_name: user-facing name for the IdP
idp_icon: if specified in the IdP config, an MXC URI for an icon
for the IdP
idp_brand: if specified in the IdP config, a textual identifier
for the brand of the IdP
sso_auth_success.html: HTML page shown after a successful user interactive
authentication session.
Note that this page must include the JavaScript which notifies of a successful
authentication (see https://matrix.org/docs/spec/client_server/r0.6.0#fallback).
This template has no additional variables.
sso_auth_bad_user.html: HTML page shown after a user-interactive authentication
session which does not map correctly onto the expected user.
When rendering, this template is given the following variables:
server_name: the homeserver's name.
user_id_to_verify: the MXID of the user that we are trying to
validate.
sso_account_deactivated.html: HTML page shown during single sign-on if a deactivated
user (according to Synapse's database) attempts to login.
This template has no additional variables.
sso_error.html: HTML page to display to users if something goes wrong during the
OpenID Connect authentication process.
When rendering, this template is given two variables:
error: the technical name of the error
error_description: a human-readable message for the error
Synapse supports single sign-on through the SAML, Open ID Connect or CAS protocols.
LDAP and other login methods are supported through first and third-party password
auth provider modules.
Synapse can be configured to use an OpenID Connect Provider (OP) for
authentication, instead of its own local password database.
Any OP should work with Synapse, as long as it supports the authorization code
flow. There are a few options for that:
start a local OP. Synapse has been tested with Hydra and
Dex. Note that for an OP to work, it should be served under a
secure (HTTPS) origin. A certificate signed with a self-signed, locally
trusted CA should work. In that case, start Synapse with a SSL_CERT_FILE
environment variable set to the path of the CA.
set up a SaaS OP, like Google, Auth0 or
Okta. Synapse has been tested with Auth0 and Google.
The OpenID integration in Synapse uses the
authlib library, which must be installed
as follows:
The relevant libraries are included in the Docker images and Debian packages
provided by matrix.org so no further action is needed.
If you installed Synapse into a virtualenv, run /path/to/env/bin/pip install matrix-synapse[oidc] to install the necessary dependencies.
For other installation mechanisms, see the documentation provided by the
maintainer.
To enable the OpenID integration, you should then add a section to the oidc_providers
setting in your configuration file.
See the configuration manual for some sample settings, as well as
the text below for example configurations for specific providers.
This lets the OpenID Connect Provider notify Synapse when a user logs out, so that Synapse can end that user session.
This feature can be enabled by setting the backchannel_logout_enabled property to true in the provider configuration, and setting the following URL as destination for Back-Channel Logout notifications in your OpenID Connect Provider: [synapse public baseurl]/_synapse/client/oidc/backchannel_logout
Azure AD can act as an OpenID Connect Provider. Register a new application under
App registrations in the Azure AD management console. The RedirectURI for your
application should point to your matrix server:
[synapse public baseurl]/_synapse/client/oidc/callback
Go to Certificates & secrets and register a new client secret. Make note of your
Directory (tenant) ID as it will be used in the Azure links.
Edit your Synapse config file and change the oidc_config section:
- idp_id: apple
idp_name: Apple
issuer: "https://appleid.apple.com"
client_id: "your-client-id" # Set to the "identifier" for your "ServicesID"
client_auth_method: "client_secret_post"
client_secret_jwt_key:
key_file: "/path/to/AuthKey_KEYIDCODE.p8" # point to your key file
jwt_header:
alg: ES256
kid: "KEYIDCODE" # Set to the 10-char Key ID
jwt_payload:
iss: TEAMIDCODE # Set to the 10-char Team ID
scopes: ["name", "email", "openid"]
authorization_endpoint: https://appleid.apple.com/auth/authorize?response_mode=form_post
user_mapping_provider:
config:
email_template: "{{ user.email }}"
Set the Allowed Callback URLs to [synapse public baseurl]/_synapse/client/oidc/callback
Add a rule with any name to add the preferred_username claim.
(See https://auth0.com/docs/customize/rules/create-rules for more information on how to create rules.)
Create a provider in Authentik, with type OAuth2/OpenID.
The parameters are:
Client Type: Confidential
JWT Algorithm: RS256
Scopes: OpenID, Email and Profile
RSA Key: Select any available key
Redirect URIs: [synapse public baseurl]/_synapse/client/oidc/callback
Create an application for synapse in Authentik and link it to the provider.
Note the slug of your application, Client ID and Client Secret.
Note: RSA keys must be used for signing for Authentik, ECC keys do not work.
Synapse config:
oidc_providers:
- idp_id: authentik
idp_name: authentik
discover: true
issuer: "https://your.authentik.example.org/application/o/your-app-slug/" # TO BE FILLED: domain and slug
client_id: "your client id" # TO BE FILLED
client_secret: "your client secret" # TO BE FILLED
scopes:
- "openid"
- "profile"
- "email"
user_mapping_provider:
config:
localpart_template: "{{ user.preferred_username }}"
display_name_template: "{{ user.preferred_username|capitalize }}" # TO BE FILLED: If your users have names in Authentik and you want those in Synapse, this should be replaced with user.name|capitalize.
Dex is a simple, open-source OpenID Connect Provider.
Although it is designed to help building a full-blown provider with an
external database, it can be configured with static passwords in a config file.
django-oauth-toolkit is a
Django application providing out of the box all the endpoints, data and logic
needed to add OAuth2 capabilities to your Django projects. It supports
OpenID Connect too.
Configuration on Django's side:
Add an application: https://example.com/admin/oauth2_provider/application/add/ and choose parameters like this:
Facebook do have an OIDC discovery endpoint,
but it has a response_types_supported which excludes "code" (which we rely on, and
is even mentioned in their documentation),
so we have to disable discovery and configure the URIs manually.
GitHub is a bit special as it is not an OpenID Connect compliant provider, but
just a regular OAuth2 provider.
The /user API endpoint
can be used to retrieve information on the authenticated user. As the Synapse
login mechanism needs an attribute to uniquely identify users, and that endpoint
does not return a sub property, an alternative subject_claim has to be set.
Gitea is, like Github, not an OpenID provider, but just an OAuth2 provider.
The /user API endpoint
can be used to retrieve information on the authenticated user. As the Synapse
login mechanism needs an attribute to uniquely identify users, and that endpoint
does not return a sub property, an alternative subject_claim has to be set.
Create a new application.
Add this Callback URL: [synapse public baseurl]/_synapse/client/oidc/callback
Synapse config:
oidc_providers:
- idp_id: gitea
idp_name: Gitea
discover: false
issuer: "https://your-gitea.com/"
client_id: "your-client-id" # TO BE FILLED
client_secret: "your-client-secret" # TO BE FILLED
client_auth_method: client_secret_post
scopes: [] # Gitea doesn't support Scopes
authorization_endpoint: "https://your-gitea.com/login/oauth/authorize"
token_endpoint: "https://your-gitea.com/login/oauth/access_token"
userinfo_endpoint: "https://your-gitea.com/api/v1/user"
user_mapping_provider:
config:
subject_claim: "id"
localpart_template: "{{ user.login }}"
display_name_template: "{{ user.full_name }}"
Keycloak is an opensource IdP maintained by Red Hat.
Keycloak supports OIDC Back-Channel Logout, which sends logout notification to Synapse, so that Synapse users get logged out when they log out from Keycloak.
This can be optionally enabled by setting backchannel_logout_enabled to true in the Synapse configuration, and by setting the "Backchannel Logout URL" in Keycloak.
Create an OpenID Connect Relying Parties in LemonLDAP::NG
The parameters are:
Client ID under the basic menu of the new Relying Parties (Options > Basic > Client ID)
Client secret (Options > Basic > Client secret)
JWT Algorithm: RS256 within the security menu of the new Relying Parties
(Options > Security > ID Token signature algorithm and Options > Security > Access Token signature algorithm)
Scopes: OpenID, Email and Profile
Allowed redirection addresses for login (Options > Basic > Allowed redirection addresses for login ) :
[synapse public baseurl]/_synapse/client/oidc/callback
Synapse config:
oidc_providers:
- idp_id: lemonldap
idp_name: lemonldap
discover: true
issuer: "https://auth.example.org/" # TO BE FILLED: replace with your domain
client_id: "your client id" # TO BE FILLED
client_secret: "your client secret" # TO BE FILLED
scopes:
- "openid"
- "profile"
- "email"
user_mapping_provider:
config:
localpart_template: "{{ user.preferred_username }}}"
# TO BE FILLED: If your users have names in LemonLDAP::NG and you want those in Synapse, this should be replaced with user.name|capitalize or any valid filter.
display_name_template: "{{ user.preferred_username|capitalize }}"
Mastodon instances provide an OAuth API, allowing those instances to be used as a single sign-on provider for Synapse.
The first step is to register Synapse as an application with your Mastodon instance, using the Create an application API (see also here). There are several ways to do this, but in the example below we are using CURL.
This example assumes that:
the Mastodon instance website URL is https://your.mastodon.instance.url, and
Synapse will be registered as an app named my_synapse_app.
Send the following request, substituting the value of synapse_public_baseurl from your Synapse installation.
curl -d "client_name=my_synapse_app&redirect_uris=https://[synapse_public_baseurl]/_synapse/client/oidc/callback" -X POST https://your.mastodon.instance.url/api/v1/apps
You should receive a response similar to the following. Make sure to save it.
As the Synapse login mechanism needs an attribute to uniquely identify users, and Mastodon's endpoint does not return a sub property, an alternative subject_template has to be set. Your Synapse configuration should include the following:
A mapping provider is a Python class (loaded via a Python module) that
works out how to map attributes of a SSO response to Matrix-specific
user attributes. Details such as user ID localpart, displayname, and even avatar
URLs are all things that can be mapped from talking to a SSO service.
As an example, a SSO service may return the email address
"john.smith@example.com" for a user, whereas Synapse will need to figure out how
to turn that into a displayname when creating a Matrix user for this individual.
It may choose John Smith, or Smith, John [Example.com] or any number of
variations. As each Synapse configuration may want something different, this is
where SAML mapping providers come into play.
SSO mapping providers are currently supported for OpenID and SAML SSO
configurations. Please see the details below for how to implement your own.
It is up to the mapping provider whether the user should be assigned a predefined
Matrix ID based on the SSO attributes, or if the user should be allowed to
choose their own username.
In the first case - where users are automatically allocated a Matrix ID - it is
the responsibility of the mapping provider to normalise the SSO attributes and
map them to a valid Matrix ID. The specification for Matrix
IDs has some
information about what is considered valid.
If the mapping provider does not assign a Matrix ID, then Synapse will
automatically serve an HTML page allowing the user to pick their own username.
External mapping providers are provided to Synapse in the form of an external
Python module. You can retrieve this module from PyPI or elsewhere,
but it must be importable via Synapse (e.g. it must be in the same virtualenv
as Synapse). The Synapse config is then modified to point to the mapping provider
(and optionally provide additional configuration for it).
oidc_providers.user_mapping_provider.config allows you to provide custom
configuration options to the module. Check with the module's documentation for
what options it provides (if any). The options listed by default are for the
user mapping provider built in to Synapse. If using a custom module, you should
comment these options out and use those specified by the module instead.
A custom mapping provider must specify the following methods:
def __init__(self, parsed_config)
Arguments:
parsed_config - A configuration object that is the return value of the
parse_config method. You should set any configuration options needed by
the module here.
def parse_config(config)
This method should have the @staticmethod decoration.
Arguments:
config - A dict representing the parsed content of the
oidc_providers.user_mapping_provider.config homeserver config option.
Runs on homeserver startup. Providers should extract and validate
any option values they need here.
Whatever is returned will be passed back to the user mapping provider module's
__init__ method during construction.
def get_remote_user_id(self, userinfo)
Arguments:
userinfo - A authlib.oidc.core.claims.UserInfo object to extract user
information from.
This method must return a string, which is the unique, immutable identifier
for the user. Commonly the sub claim of the response.
token - A dictionary which includes information necessary to make
further requests to the OpenID provider.
failures - An int that represents the amount of times the returned
mxid localpart mapping has failed. This should be used
to create a deduplicated mxid localpart which should be
returned instead. For example, if this method returns
john.doe as the value of localpart in the returned
dict, and that is already taken on the homeserver, this
method will be called again with the same parameters but
with failures=1. The method should then return a different
localpart value, such as john.doe1.
Returns a dictionary with two keys:
localpart: A string, used to generate the Matrix ID. If this is
None, the user is prompted to pick their own username. This is only used
during a user's first login. Once a localpart has been associated with a
remote user ID (see get_remote_user_id) it cannot be updated.
confirm_localpart: A boolean. If set to True, when a localpart
string is returned from this method, Synapse will prompt the user to
either accept this localpart or pick their own username. Otherwise this
option has no effect. If omitted, defaults to False.
display_name: An optional string, the display name for the user.
emails: A list of strings, the email address(es) to associate with
this user. If omitted, defaults to an empty list.
saml2_config.user_mapping_provider.config allows you to provide custom
configuration options to the module. Check with the module's documentation for
what options it provides (if any). The options listed by default are for the
user mapping provider built in to Synapse. If using a custom module, you should
comment these options out and use those specified by the module instead.
A custom mapping provider must specify the following methods:
def __init__(self, parsed_config, module_api)
Arguments:
parsed_config - A configuration object that is the return value of the
parse_config method. You should set any configuration options needed by
the module here.
module_api - a synapse.module_api.ModuleApi object which provides the
stable API available for extension modules.
def parse_config(config)
This method should have the @staticmethod decoration.
Arguments:
config - A dict representing the parsed content of the
saml_config.user_mapping_provider.config homeserver config option.
Runs on homeserver startup. Providers should extract and validate
any option values they need here.
Whatever is returned will be passed back to the user mapping provider module's
__init__ method during construction.
def get_saml_attributes(config)
This method should have the @staticmethod decoration.
Arguments:
config - A object resulting from a call to parse_config.
Returns a tuple of two sets. The first set equates to the SAML auth
response attributes that are required for the module to function, whereas
the second set consists of those attributes which can be used if available,
but are not necessary.
saml_response - A saml2.response.AuthnResponse object to extract user
information from.
failures - An int that represents the amount of times the returned
mxid localpart mapping has failed. This should be used
to create a deduplicated mxid localpart which should be
returned instead. For example, if this method returns
john.doe as the value of mxid_localpart in the returned
dict, and that is already taken on the homeserver, this
method will be called again with the same parameters but
with failures=1. The method should then return a different
mxid_localpart value, such as john.doe1.
client_redirect_url - A string, the URL that the client will be
redirected to.
This method must return a dictionary, which will then be used by Synapse
to build a new user. The following keys are allowed:
mxid_localpart - A string, the mxid localpart of the new user. If this is
None, the user is prompted to pick their own username. This is only used
during a user's first login. Once a localpart has been associated with a
remote user ID (see get_remote_user_id) it cannot be updated.
displayname - The displayname of the new user. If not provided, will default to
the value of mxid_localpart.
emails - A list of emails for the new user. If not provided, will
default to an empty list.
Alternatively it can raise a synapse.api.errors.RedirectException to
redirect the user to another page. This is useful to prompt the user for
additional information, e.g. if you want them to provide their own username.
It is the responsibility of the mapping provider to either redirect back
to client_redirect_url (including any additional information) or to
complete registration using methods from the ModuleApi.
This page of the Synapse documentation is now deprecated. For up to date
documentation on setting up or writing a password auth provider module, please see
this page.
Password auth providers offer a way for server administrators to
integrate their Synapse installation with an existing authentication
system.
A password auth provider is a Python class which is dynamically loaded
into Synapse, and provides a number of methods by which it can integrate
with the authentication system.
This document serves as a reference for those looking to implement their
own password auth providers. Additionally, here is a list of known
password auth provider module implementations:
Password auth provider classes must provide the following methods:
parse_config(config)
This method is passed the config object for this module from the
homeserver configuration file.
It should perform any appropriate sanity checks on the provided
configuration, and return an object which is then passed into
__init__.
This method should have the @staticmethod decoration.
__init__(self, config, account_handler)
The constructor is passed the config object returned by
parse_config, and a synapse.module_api.ModuleApi object which
allows the password provider to check if accounts exist and/or create
new ones.
Password auth provider classes may optionally provide the following methods:
get_db_schema_files(self)
This method, if implemented, should return an Iterable of
(name, stream) pairs of database schema files. Each file is applied
in turn at initialisation, and a record is then made in the database
so that it is not re-applied on the next start.
get_supported_login_types(self)
This method, if implemented, should return a dict mapping from a
login type identifier (such as m.login.password) to an iterable
giving the fields which must be provided by the user in the submission
to the /login API.
These fields are passed in the login_dict dictionary to check_auth.
For example, if a password auth provider wants to implement a custom
login type of com.example.custom_login, where the client is expected
to pass the fields secret1 and secret2, the provider should
implement this method and return the following dict:
This method does the real work. If implemented, it
will be called for each login attempt where the login type matches one
of the keys returned by get_supported_login_types.
It is passed the (possibly unqualified) user field provided by the client,
the login type, and a dictionary of login secrets passed by the
client.
The method should return an Awaitable object, which resolves
to the canonical @localpart:domain user ID if authentication is
successful, and None if not.
Alternatively, the Awaitable can resolve to a (str, func) tuple, in
which case the second field is a callback which will be called with
the result from the /login call (including access_token,
device_id, etc.)
check_3pid_auth(self, medium, address, password)
This method, if implemented, is called when a user attempts to
register or log in with a third party identifier, such as email. It is
passed the medium (ex. "email"), an address (ex.
"jdoe@example.com") and the user's password.
The method should return an Awaitable object, which resolves
to a str containing the user's (canonical) User id if
authentication was successful, and None if not.
As with check_auth, the Awaitable may alternatively resolve to a
(user_id, callback) tuple.
check_password(self, user_id, password)
This method provides a simpler interface than
get_supported_login_types and check_auth for password auth
providers that just want to provide a mechanism for validating
m.login.password logins.
If implemented, it will be called to check logins with an
m.login.password login type. It is passed a qualified
@localpart:domain user id, and the password provided by the user.
The method should return an Awaitable object, which resolves
to True if authentication is successful, and False if not.
This method, if implemented, is called when a user logs out. It is
passed the qualified user ID, the ID of the deactivated device (if
any: access tokens are occasionally created without an associated
device ID), and the (now deactivated) access token.
It may return an Awaitable object; the logout request will
wait for the Awaitable to complete, but the result is ignored.
Synapse comes with a non-standard login type to support
JSON Web Tokens. In general the
documentation for
the login endpoint
is still valid (and the mechanism works similarly to the
token based login).
To log in using a JSON Web Token, clients should submit a /login request as
follows:
The token field should include the JSON web token with the following claims:
A claim that encodes the local part of the user ID is required. By default,
the sub (subject) claim is used, or a custom claim can be set in the
configuration file.
The expiration time (exp), not before time (nbf), and issued at (iat)
claims are optional, but validated if present.
The issuer (iss) claim is optional, but required and validated if configured.
The audience (aud) claim is optional, but required and validated if configured.
Providing the audience claim when not configured will cause validation to fail.
In the case that the token is not valid, the homeserver must respond with
403 Forbidden and an error code of M_FORBIDDEN.
As with other login types, there are additional fields (e.g. device_id and
initial_device_display_name) which can be included in the above request.
The JSON Web Token integration in Synapse uses the
Authlib library, which must be installed
as follows:
The relevant libraries are included in the Docker images and Debian packages
provided by matrix.org so no further action is needed.
If you installed Synapse into a virtualenv, run /path/to/env/bin/pip install synapse[jwt] to install the necessary dependencies.
For other installation mechanisms, see the documentation provided by the
maintainer.
To enable the JSON web token integration, you should then add a jwt_config option
to your configuration file. See the configuration manual for some
sample settings.
You can use the following short Python snippet to generate a JWT
protected by an HMAC.
Take care that the secret and the algorithm given in the header match
the entries from jwt_config above.
Synapse supports refresh tokens since version 1.49 (some earlier versions had support for an earlier, experimental draft of MSC2918 which is not compatible).
Synapse users' sessions are identified by access tokens; access tokens are
issued to users on login. Each session gets a unique access token which identifies
it; the access token must be kept secret as it grants access to the user's account.
Traditionally, these access tokens were eternally valid (at least until the user
explicitly chose to log out).
In some cases, it may be desirable for these access tokens to expire so that the
potential damage caused by leaking an access token is reduced.
On the other hand, forcing a user to re-authenticate (log in again) often might
be too much of an inconvenience.
Refresh tokens are a mechanism to avoid some of this inconvenience whilst
still getting most of the benefits of short access token lifetimes.
Refresh tokens are also a concept present in OAuth 2 — further reading is available
here.
When refresh tokens are in use, both an access token and a refresh token will be
issued to users on login. The access token will expire after a predetermined amount
of time, but otherwise works in the same way as before. When the access token is
close to expiring (or has expired), the user's client should present the homeserver
(Synapse) with the refresh token.
The homeserver will then generate a new access token and refresh token for the user
and return them. The old refresh token is invalidated and can not be used again*.
Finally, refresh tokens also make it possible for sessions to be logged out if they
are inactive for too long, before the session naturally ends; see the configuration
guide below.
*To prevent issues if clients lose connection half-way through refreshing a token,
the refresh token is only invalidated once the new access token has been used at
least once. For all intents and purposes, the above simplification is sufficient.
If a third party gets both your access token and refresh token, they will be able to
continue to enjoy access to your session.
This is still an improvement because you (the user) will notice when your
session expires and you're not able to use your refresh token.
That would be a giveaway that someone else has compromised your session.
You would be able to log in again and terminate that session.
Previously (with long-lived access tokens), a third party that has your access
token could go undetected for a very long time.
Clients need to implement support for refresh tokens in order for them to be a
useful mechanism.
It is up to homeserver administrators if they want to issue long-lived access
tokens to clients not implementing refresh tokens.
For compatibility, it is likely that they should, at least until client support
is widespread.
Users with clients that support refresh tokens will still benefit from the
added security; it's not possible to downgrade a session to using long-lived
access tokens so this effectively gives users the choice.
In a closed environment where all users use known clients, this may not be
an issue as the homeserver administrator can know if the clients have refresh
token support. In that case, the non-refreshable access token lifetime
may be set to a short duration so that a similar level of security is provided.
The following configuration options, in the registration section, are related:
session_lifetime: maximum length of a session, even if it's refreshed.
In other words, the client must log in again after this time period.
In most cases, this can be unset (infinite) or set to a long time (years or months).
refreshable_access_token_lifetime: lifetime of access tokens that are created
by clients supporting refresh tokens.
This should be short; a good value might be 5 minutes (5m).
nonrefreshable_access_token_lifetime: lifetime of access tokens that are created
by clients which don't support refresh tokens.
Make this short if you want to effectively force use of refresh tokens.
Make this long if you don't want to inconvenience users of clients which don't
support refresh tokens (by forcing them to frequently re-authenticate using
login credentials).
refresh_token_lifetime: lifetime of refresh tokens.
In other words, the client must refresh within this time period to maintain its session.
Unless you want to log inactive sessions out, it is often fine to use a long
value here or even leave it unset (infinite).
Beware that making it too short will inconvenience clients that do not connect
very often, including mobile clients and clients of infrequent users (by making
it more difficult for them to refresh in time, which may force them to need to
re-authenticate using login credentials).
Note: All four options above only apply when tokens are created (by logging in or refreshing).
Changes to these settings do not apply retroactively.
If you'd like to force sessions to be logged out upon inactivity, you can enable
refreshable access token expiry and refresh token expiry.
This works because a client must refresh at least once within a period of
refresh_token_lifetime in order to maintain valid credentials to access the
account.
(It's suggested that refresh_token_lifetime should be longer than
refreshable_access_token_lifetime and this section assumes that to be the case
for simplicity.)
Note: this will only affect sessions using refresh tokens. You may wish to
set a short nonrefreshable_access_token_lifetime to prevent this being bypassed
by clients that do not support refresh tokens.
It may be desirable to permit some short periods of inactivity, for example to
accommodate brief outages in client connectivity.
The following model aims to provide guidance for choosing refresh_token_lifetime
and refreshable_access_token_lifetime to satisfy requirements of the form:
inactivity longer than LMUST cause the session to be logged out; and
inactivity shorter than SMUST NOT cause the session to be logged out.
This model makes the weakest assumption that all active clients will refresh as
needed to maintain an active access token, but no sooner.
In reality, clients may refresh more often than this model assumes, but the
above requirements will still hold.
To satisfy the above model,
refresh_token_lifetime should be set to L; and
refreshable_access_token_lifetime should be set to L - S.
A captcha can be enabled on your homeserver to help prevent bots from registering
accounts. Synapse currently uses Google's reCAPTCHA service which requires API keys
from Google.
Set the type to reCAPTCHA v2 using the "I'm not a robot" Checkbox option.
This is the only type of captcha that works with Synapse.
Add the public hostname for your server, as set in public_baseurl
in homeserver.yaml, to the list of authorized domains. If you have not set
public_baseurl, use server_name.
Agree to the terms of service and submit.
Copy your site key and secret key and add them to your homeserver.yaml
configuration file
Go to the settings page for the CAPTCHA you just created
Uncheck the "Verify the origin of reCAPTCHA solutions" checkbox so that the
captcha can be displayed in any client. If you do not disable this option then you
must specify the domains of every client that is allowed to display the CAPTCHA.
The reCAPTCHA API requires that the IP address of the user who solved the
CAPTCHA is sent. If the client is connecting through a proxy or load balancer,
it may be required to use the X-Forwarded-For (XFF) header instead of the origin
IP address. This can be configured using the x_forwarded directive in the
listeners section of the homeserver.yaml configuration file.
The registration of new application services depends on the homeserver used.
In synapse, you need to create a new configuration file for your AS and add it
to the list specified under the app_service_config_files config
option in your synapse config.
The format of the AS configuration file is as follows:
id: <your-AS-id>
url: <base url of AS>
as_token: <token AS will add to requests to HS>
hs_token: <token HS will add to requests to AS>
sender_localpart: <localpart of AS user>
namespaces:
users: # List of users we're interested in
- exclusive: <bool>
regex: <regex>
group_id: <group>
- ...
aliases: [] # List of aliases we're interested in
rooms: [] # List of room ids we're interested in
exclusive: If enabled, only this application service is allowed to register users in its namespace(s).
group_id: All users of this application service are dynamically joined to this group. This is useful for e.g user organisation or flairs.
See the spec for further details on how application services work.
'Server Notices' are a new feature introduced in Synapse 0.30. They provide a
channel whereby server administrators can send messages to users on the server.
They are used as part of communication of the server polices (see
Consent Tracking), however the intention is that
they may also find a use for features such as "Message of the day".
This is a feature specific to Synapse, but it uses standard Matrix
communication mechanisms, so should work with any Matrix client.
When the user is first sent a server notice, they will get an invitation to a
room (typically called 'Server Notices', though this is configurable in
homeserver.yaml). They will be unable to reject this invitation -
attempts to do so will receive an error.
Once they accept the invitation, they will see the notice message in the room
history; it will appear to have come from the 'server notices user' (see
below).
The user is prevented from sending any messages in this room by the power
levels.
Having joined the room, the user can leave the room if they want. Subsequent
server notices will then cause a new room to be created.
Server notices come from a specific user id on the server. Server
administrators are free to choose the user id - something like server is
suggested, meaning the notices will come from
@server:<your_server_name>. Once the Server Notices user is configured, that
user id becomes a special, privileged user, so administrators should ensure
that it is not already allocated.
In order to support server notices, it is necessary to add some configuration
to the homeserver.yaml file. In particular, you should add a server_notices
section, which should look like this:
The only compulsory setting is system_mxid_localpart, which defines the user
id of the Server Notices user, as above. room_name defines the name of the
room which will be created.
system_mxid_display_name and system_mxid_avatar_url can be used to set the
displayname and avatar of the Server Notices user.
auto_join will autojoin users to the notices room instead of sending an invite.
Synapse 0.30 introduces support for tracking whether users have agreed to the
terms and conditions set by the administrator of a server - and blocking access
to the server until they have.
There are several parts to this functionality; each requires some specific
configuration in homeserver.yaml to be enabled.
Note that various parts of the configuration and this document refer to the
"privacy policy": agreement with a privacy policy is one particular use of this
feature, but of course administrators can specify other terms and conditions
unrelated to "privacy" per se.
Synapse can be configured to serve the user a simple policy form with an
"accept" button. Clicking "Accept" records the user's acceptance in the
database and shows a success page.
To enable this, first create templates for the policy and success pages.
These should be stored on the local filesystem.
Note that the templates must be stored under a name giving the language of the
template - currently this must always be en (for "English");
internationalisation support is intended for the future.
The template for the policy itself should be versioned and named according to
the version: for example 1.0.html. The version of the policy which the user
has agreed to is stored in the database.
Once the templates are in place, make the following changes to homeserver.yaml:
Add a user_consent section, which should look like:
template_dir points to the directory containing the policy
templates. version defines the version of the policy which will be served
to the user. In the example above, Synapse will serve
privacy_policy_templates/en/1.0.html.
Add a form_secret setting at the top level:
form_secret: "<unique secret>"
This should be set to an arbitrary secret string (try pwgen -y 30 to
generate suitable secrets).
More on what this is used for below.
Add consent wherever the client resource is currently enabled in the
listeners configuration. For example:
Finally, ensure that jinja2 is installed. If you are using a virtualenv, this
should be a matter of pip install Jinja2. On debian, try apt-get install python-jinja2.
Once this is complete, and the server has been restarted, try visiting
https://<server>/_matrix/consent. If correctly configured, this should give
an error "Missing string query parameter 'u'". It is now possible to manually
construct URIs where users can give their consent.
user_consent:
require_at_registration: true
policy_name: "Privacy Policy" # or whatever you'd like to call the policy
In your consent templates, make use of the public_version variable to
see if an unauthenticated user is viewing the page. This is typically
wrapped around the form that would be used to actually agree to the document:
{% if not public_version %}
<!-- The variables used here are only provided when the 'u' param is given to the homeserver -->
<form method="post" action="consent">
<input type="hidden" name="v" value="{{version}}"/>
<input type="hidden" name="u" value="{{user}}"/>
<input type="hidden" name="h" value="{{userhmac}}"/>
<input type="submit" value="Sure thing!"/>
</form>
{% endif %}
Restart Synapse to apply the changes.
Visiting https://<server>/_matrix/consent should now give you a view of the privacy
document. This is what users will be able to see when registering for accounts.
It may be useful to manually construct the "consent URI" for a given user - for
instance, in order to send them an email asking them to consent. To do this,
take the base https://<server>/_matrix/consent URL and add the following
query parameters:
u: the user id of the user. This can either be a full MXID
(@user:server.com) or just the localpart (user).
h: hex-encoded HMAC-SHA256 of u using the form_secret as a key. It is
possible to calculate this on the commandline with something like:
This should result in a URI which looks something like:
https://<server>/_matrix/consent?u=<user>&h=68a152465a4d....
Note that not providing a u parameter will be interpreted as wanting to view
the document from an unauthenticated perspective, such as prior to registration.
Therefore, the h parameter is not required in this scenario. To enable this
behaviour, set require_at_registration to true in your user_consent config.
Add server_notice_content under user_consent in homeserver.yaml. For
example:
user_consent:
server_notice_content:
msgtype: m.text
body: >-
Please give your consent to the privacy policy at %(consent_uri)s.
Synapse automatically replaces the placeholder %(consent_uri)s with the
consent uri for that user.
ensure that public_baseurl is set in homeserver.yaml, and gives the base
URI that clients use to connect to the server. (It is used to construct
consent_uri in the server notice.)
Synapse can be configured to block any attempts to join rooms or send messages
until the user has given their agreement to the policy. (Joining the server
notices room is exempted from this).
To enable this, add block_events_error under user_consent. For example:
user_consent:
block_events_error: >-
You can't send any messages until you consent to the privacy policy at
%(consent_uri)s.
Synapse automatically replaces the placeholder %(consent_uri)s with the
consent uri for that user.
ensure that public_baseurl is set in homeserver.yaml, and gives the base
URI that clients use to connect to the server. (It is used to construct
consent_uri in the error.)
The user directory is maintained based on users that are 'visible' to the homeserver -
i.e. ones which are local to the server and ones which any local user shares a
room with.
The directory info is stored in various tables, which can sometimes get out of
sync (although this is considered a bug). If this happens, for now the
solution to fix it is to use the admin API
and execute the job regenerate_directory. This should then start a background task to
flush the current tables and regenerate the directory. Depending on the size
of your homeserver (number of users and rooms) this can take a while.
There are five relevant tables that collectively form the "user directory".
Three of them track a list of all known users. The last two (collectively called
the "search tables") track which users are visible to each other.
From all of these tables we exclude three types of local user:
support users
appservice users
deactivated users
A description of each table follows:
user_directory. This contains the user ID, display name and avatar of each user.
Because there is only one directory entry per user, it is important that it
only contain publicly visible information. Otherwise, this will leak the
nickname or avatar used in a private room.
Indexed on rooms. Indexed on users.
user_directory_search. To be joined to user_directory. It contains an extra
column that enables full text search based on user IDs and display names.
Different schemas for SQLite and Postgres are used.
Indexed on the full text search data. Indexed on users.
user_directory_stream_pos. When the initial background update to populate
the directory is complete, we record a stream position here. This indicates
that synapse should now listen for room changes and incrementally update
the directory where necessary. (See stream positions.)
users_in_public_rooms. Contains associations between users and the public
rooms they're in. Used to determine which users are in public rooms and should
be publicly visible in the directory. Both local and remote users are tracked.
users_who_share_private_rooms. Rows are triples (L, M, room id) where L
is a local user and M is a local or remote user. L and M should be
different, but this isn't enforced by a constraint.
Note that if two local users share a room then there will be two entries:
(user1, user2, !room_id) and (user2, user1, !room_id).
If search_all_users is false, then results are limited to users who:
Are found in the users_in_public_rooms table, or
Are found in the users_who_share_private_rooms where L is the requesting
user and M is the search result.
Otherwise, if search_all_users is true, no such limits are placed and all
users known to the server (matching the search query) will be returned.
By default, locked users are not returned. If show_locked_users is true then
no filtering on the locked status of a user is done.
The user provided search term is lowercased and normalized using NFKC,
this treats the string as case-insensitive, canonicalizes different forms of the
same text, and maps some "roughly equivalent" characters together.
If unavailable, then runs of ASCII characters, numbers, underscores, and hyphens
are considered words.
The queries for PostgreSQL and SQLite are detailed below, by their overall goal
is to find matching users, preferring users who are "real" (e.g. not bots,
not deactivated). It is assumed that real users will have an display name and
avatar set.
The above words are then transformed into two queries:
"exact" which matches the parsed words exactly (using to_tsquery);
"prefix" which matches the parsed words as prefixes (using to_tsquery).
Results are composed of all rows in the user_directory_search table whose information
matches one (or both) of these queries. Results are ordered by calculating a weighted
score for each result, higher scores are returned first:
4x if a user ID exists.
1.2x if the user has a display name set.
1.2x if the user has an avatar set.
0x-3x by the full text search results using the ts_rank_cd function
against the "exact" search query; this has four variables with the following weightings:
D: 0.1 for the user ID's domain
C: 0.1 for unused
B: 0.9 for the user's display name (or an empty string if it is not set)
A: 0.1 for the user ID's localpart
0x-1x by the full text search results using the ts_rank_cd function against the
"prefix" search query. (Using the same weightings as above.)
If prefer_local_users is true, then 2x if the user is local to the homeserver.
Note that ts_rank_cd returns a weight between 0 and 1. The initial weighting of
all results is 1.
Results are composed of all rows in the user_directory_search whose information
matches the query. Results are ordered by the following information, with each
subsequent column used as a tiebreaker, for each result:
By the rank
of the full text search results using the matchinfo function. Higher
ranks are returned first.
If prefer_local_users is true, then users local to the homeserver are
returned first.
Synapse admins can enable support for message retention policies on
their homeserver. Message retention policies exist at a room level,
follow the semantics described in
MSC1763,
and allow server and room admins to configure how long messages should
be kept in a homeserver's database before being purged from it.
Please note that, as this feature isn't part of the Matrix
specification yet, this implementation is to be considered as
experimental.
A message retention policy is mainly defined by its max_lifetime
parameter, which defines how long a message can be kept around after
it was sent to the room. If a room doesn't have a message retention
policy, and there's no default one for a given server, then no message
sent in that room is ever purged on that server.
MSC1763 also specifies semantics for a min_lifetime parameter which
defines the amount of time after which an event can get purged (after
it was sent to the room), but Synapse doesn't currently support it
beyond registering it.
Both max_lifetime and min_lifetime are optional parameters.
Note that message retention policies don't apply to state events.
Once an event reaches its expiry date (defined as the time it was sent
plus the value for max_lifetime in the room), two things happen:
Synapse stops serving the event to clients via any endpoint.
The message gets picked up by the next purge job (see the "Purge jobs"
section) and is removed from Synapse's database.
Since purge jobs don't run continuously, this means that an event might
stay in a server's database for longer than the value for max_lifetime
in the room would allow, though hidden from clients.
Similarly, if a server (with support for message retention policies
enabled) receives from another server an event that should have been
purged according to its room's policy, then the receiving server will
process and store that event until it's picked up by the next purge job,
though it will always hide it from clients.
Synapse requires at least one message in each room, so it will never
delete the last message in a room. It will, however, hide it from
clients.
A default message retention policy is a policy defined in Synapse's
configuration that is used by Synapse for every room that doesn't have a
message retention policy configured in its state. This allows server
admins to ensure that messages are never kept indefinitely in a server's
database.
A default policy can be defined as such, by adding the retention option in
the configuration file and adding these sub-options:
default_policy:
min_lifetime: 1d
max_lifetime: 1y
Here, min_lifetime and max_lifetime have the same meaning and level
of support as previously described. They can be expressed either as a
duration (using the units s (seconds), m (minutes), h (hours),
d (days), w (weeks) and y (years)) or as a number of milliseconds.
Purge jobs are the jobs that Synapse runs in the background to purge
expired events from the database. They are only run if support for
message retention policies is enabled in the server's configuration. If
no configuration for purge jobs is configured by the server admin,
Synapse will use a default configuration, which is described here in the
configuration manual.
Some server admins might want a finer control on when events are removed
depending on an event's room's policy. This can be done by setting the
purge_jobs sub-section in the retention section of the configuration
file. An example of such configuration could be:
purge_jobs:
- longest_max_lifetime: 3d
interval: 12h
- shortest_max_lifetime: 3d
longest_max_lifetime: 1w
interval: 1d
- shortest_max_lifetime: 1w
interval: 2d
In this example, we define three jobs:
one that runs twice a day (every 12 hours) and purges events in rooms
which policy's max_lifetime is lower or equal to 3 days.
one that runs once a day and purges events in rooms which policy's
max_lifetime is between 3 days and a week.
one that runs once every 2 days and purges events in rooms which
policy's max_lifetime is greater than a week.
Note that this example is tailored to show different configurations and
features slightly more jobs than it's probably necessary (in practice, a
server admin would probably consider it better to replace the two last
jobs with one that runs once a day and handles rooms which
policy's max_lifetime is greater than 3 days).
Keep in mind, when configuring these jobs, that a purge job can become
quite heavy on the server if it targets many rooms, therefore prefer
having jobs with a low interval that target a limited set of rooms. Also
make sure to include a job with no minimum and one with no maximum to
make sure your configuration handles every policy.
As previously mentioned in this documentation, while a purge job that
runs e.g. every day means that an expired event might stay in the
database for up to a day after its expiry, Synapse hides expired events
from clients as soon as they expire, so the event is not visible to
local users between its expiry date and the moment it gets purged from
the server's database.
Server admins can set limits on the values of max_lifetime to use when
purging old events in a room. These limits can be defined under the
retention option in the configuration file:
allowed_lifetime_min: 1d
allowed_lifetime_max: 1y
The limits are considered when running purge jobs. If necessary, the
effective value of max_lifetime will be brought between
allowed_lifetime_min and allowed_lifetime_max (inclusive).
This means that, if the value of max_lifetime defined in the room's state
is lower than allowed_lifetime_min, the value of allowed_lifetime_min
will be used instead. Likewise, if the value of max_lifetime is higher
than allowed_lifetime_max, the value of allowed_lifetime_max will be
used instead.
In the example above, we ensure Synapse never deletes events that are less
than one day old, and that it always deletes events that are over a year
old.
If a default policy is set, and its max_lifetime value is lower than
allowed_lifetime_min or higher than allowed_lifetime_max, the same
process applies.
Both parameters are optional; if one is omitted Synapse won't use it to
adjust the effective value of max_lifetime.
Like other settings in this section, these parameters can be expressed
either as a duration or as a number of milliseconds.
To configure a room's message retention policy, a room's admin or
moderator needs to send a state event in that room with the type
m.room.retention and the following content:
{
"max_lifetime": ...
}
In this event's content, the max_lifetime parameter has the same
meaning as previously described, and needs to be expressed in
milliseconds. The event's content can also include a min_lifetime
parameter, which has the same meaning and limited support as previously
described.
Note that over every server in the room, only the ones with support for
message retention policies will actually remove expired events. This
support is currently not enabled by default in Synapse.
While purge jobs actually delete data from the database, the disk space
used by the database might not decrease immediately on the database's
host. However, even though the database engine won't free up the disk
space, it will start writing new data into where the purged data was.
If you want to reclaim the freed disk space anyway and return it to the
operating system, the server admin needs to run VACUUM FULL; (or
VACUUM; for SQLite databases) on Synapse's database (see the related
PostgreSQL documentation).
Synapse supports extending its functionality by configuring external modules.
Note: When using third-party modules, you effectively allow someone else to run
custom code on your Synapse homeserver. Server admins are encouraged to verify the
provenance of the modules they use on their homeserver and make sure the modules aren't
running malicious code on their instance.
Each module is defined by a path to a Python class as well as a configuration. This
information for a given module should be available in the module's own documentation.
The order in which modules are listed in this section is important. When processing an
action that can be handled by several modules, Synapse will always prioritise the module
that appears first (i.e. is the highest in the list). This means:
If several modules register the same callback, the callback registered by the module
that appears first is used.
If several modules try to register a handler for the same HTTP path, only the handler
registered by the module that appears first is used. Handlers registered by the other
module(s) are ignored and Synapse will log a warning message about them.
Note that Synapse doesn't allow multiple modules implementing authentication checkers via
the password auth provider feature for the same login type with different fields. If this
happens, Synapse will refuse to start.
We are currently in the process of migrating module interfaces to this system. While some
interfaces might be compatible with it, others still require configuring modules in
another part of Synapse's configuration file.
Currently, only the following pre-existing interfaces are compatible with this new system:
A module is a Python class that uses Synapse's module API to interact with the
homeserver. It can register callbacks that Synapse will call on specific operations, as
well as web resources to attach to Synapse's web server.
When instantiated, a module is given its parsed configuration as well as an instance of
the synapse.module_api.ModuleApi class. The configuration is a dictionary, and is
either the output of the module's parse_config static method (see below), or the
configuration associated with the module in Synapse's configuration file.
See the documentation for the ModuleApi class
here.
If Synapse is running with other modules configured, the order each module appears in
within the modules section of the Synapse configuration file might restrict what it can
or cannot register. See this section for more
information.
On top of the rules listed in the link above, if a callback returns a value that should
cause the current operation to fail (e.g. if a callback checking an event returns with a
value that should cause the event to be denied), Synapse will fail the operation and
ignore any subsequent callbacks that should have been run after this one.
The documentation for each callback mentions how Synapse behaves when
multiple modules implement it.
A module can implement the following static method:
@staticmethod
def parse_config(config: dict) -> Any
This method is given a dictionary resulting from parsing the YAML configuration for the
module. It may modify it (for example by parsing durations expressed as strings (e.g.
"5d") into milliseconds, etc.), and return the modified dictionary. It may also verify
that the configuration is correct, and raise an instance of
synapse.module_api.errors.ConfigError if not.
The path is the full absolute path to register the resource at. For example, if you
register a resource for the path /_synapse/client/my_super_module/say_hello, Synapse
will serve it at http(s)://[HS_URL]/_synapse/client/my_super_module/say_hello. Note
that Synapse does not allow registering resources for several sub-paths in the /_matrix
namespace (such as anything under /_matrix/client for example). It is strongly
recommended that modules register their web resources under the /_synapse/client
namespace.
The provided resource is a Python class that implements Twisted's IResource
interface (such as Resource).
Only one resource can be registered for a given path. If several modules attempt to
register a resource for the same path, the module that appears first in Synapse's
configuration file takes priority.
Modules must register their web resources in their __init__ method.
Modules can use Synapse's module API to register callbacks. Callbacks are functions that
Synapse will call when performing specific actions. Callbacks must be asynchronous (unless
specified otherwise), and are split in categories. A single module may implement callbacks
from multiple categories, and is under no obligation to implement all callbacks from the
categories it registers callbacks for.
Modules can register callbacks using one of the module API's register_[...]_callbacks
methods. The callback functions are passed to these methods as keyword arguments, with
the callback name as the argument name and the function as its value. A
register_[...]_callbacks method exists for each category.
Modules can leverage Synapse's caching tools to manage their own cached functions. This
can be helpful for modules that need to repeatedly request the same data from the database
or a remote service.
Functions that need to be wrapped with a cache need to be decorated with a @cached()
decorator (which can be imported from synapse.module_api) and registered with the
ModuleApi.register_cached_function
API when initialising the module. If the module needs to invalidate an entry in a cache,
it needs to use the ModuleApi.invalidate_cache
API, with the function to invalidate the cache of and the key(s) of the entry to
invalidate.
Below is an example of a simple module using a cached function:
from typing import Any
from synapse.module_api import cached, ModuleApi
class MyModule:
def __init__(self, config: Any, api: ModuleApi):
self.api = api
# Register the cached function so Synapse knows how to correctly invalidate
# entries for it.
self.api.register_cached_function(self.get_user_from_id)
@cached()
async def get_department_for_user(self, user_id: str) -> str:
"""A function with a cache."""
# Request a department from an external service.
return await self.http_client.get_json(
"https://int.example.com/users", {"user_id": user_id)
)["department"]
async def do_something_with_users(self) -> None:
"""Calls the cached function and then invalidates an entry in its cache."""
user_id = "@alice:example.com"
# Get the user. Since get_department_for_user is wrapped with a cache,
# the return value for this user_id will be cached.
department = await self.get_department_for_user(user_id)
# Do something with `department`...
# Let's say something has changed with our user, and the entry we have for
# them in the cache is out of date, so we want to invalidate it.
await self.api.invalidate_cache(self.get_department_for_user, (user_id,))
Spam checker callbacks allow module developers to implement spam mitigation actions for
Synapse instances. Spam checker callbacks can be registered using the module API's
register_spam_checker_callbacks method.
Changed in Synapse v1.60.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean or a string is now deprecated.
Called when receiving an event from a client or via federation. The callback must return one of:
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) a non-Codesstr to reject the operation and specify an error message. Note that clients
typically will not localize the error message to the user's preferred locale.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.61.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
Called when a user is trying to join a room. The user is represented by their Matrix user ID (e.g.
@alice:example.com) and the room is represented by its Matrix ID (e.g.
!room:example.com). The module is also given a boolean to indicate whether the user
currently has a pending invite in the room.
This callback isn't called if the join is performed by a server administrator, or in the
context of a room creation.
The callback must return one of:
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
Called when processing an invitation. Both inviter and invitee are
represented by their Matrix user ID (e.g. @alice:example.com).
The callback must return one of:
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
Called when processing an invitation using a third-party identifier (also called a 3PID,
e.g. an email address or a phone number).
The inviter is represented by their Matrix user ID (e.g. @alice:example.com), and the
invitee is represented by its medium (e.g. "email") and its address
(e.g. alice@example.com). See the Matrix specification
for more information regarding third-party identifiers.
For example, a call to this callback to send an invitation to the email address
alice@example.com would look like this:
await user_may_send_3pid_invite(
"@bob:example.com", # The inviter's user ID
"email", # The medium of the 3PID to invite
"alice@example.com", # The address of the 3PID to invite
"!some_room:example.com", # The ID of the room to send the invite into
)
Note: If the third-party identifier is already associated with a matrix user ID,
user_may_invite will be used instead.
The callback must return one of:
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
Called when trying to associate an alias with an existing room.
The callback must return one of:
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
Called when trying to publish a room to the homeserver's public rooms directory.
The callback must return one of:
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Called when computing search results in the user directory. The module must return a
bool indicating whether the given user should be excluded from user directory
searches. Return True to indicate that the user is spammy and exclude them from
search results; otherwise return False.
The profile is represented as a dictionary with the following keys:
user_id: str. The Matrix ID for this user.
display_name: Optional[str]. The user's display name, or None if this user
has not set a display name.
avatar_url: Optional[str]. The mxc:// URL to the user's avatar, or None
if this user has not set an avatar.
The module is given a copy of the original dictionary, so modifying it from within the
module cannot modify a user's profile when included in user directory search results.
If multiple modules implement this callback, they will be considered in order. If a
callback returns False, Synapse falls through to the next one. The value of the first
callback that does not return False will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called when registering a new user. The module must return a RegistrationBehaviour
indicating whether the registration can go through or must be denied, or whether the user
may be allowed to register but will be shadow banned.
The arguments passed to this callback are:
email_threepid: The email address used for registering, if any.
username: The username the user would like to register. Can be None, meaning that
Synapse will generate one later.
request_info: A collection of tuples, which first item is a user agent, and which
second item is an IP address. These user agents and IP addresses are the ones that were
used during the registration process.
auth_provider_id: The identifier of the SSO authentication provider, if any.
If multiple modules implement this callback, they will be considered in order. If a
callback returns RegistrationBehaviour.ALLOW, Synapse falls through to the next one.
The value of the first callback that does not return RegistrationBehaviour.ALLOW will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Changed in Synapse v1.62.0: synapse.module_api.NOT_SPAM and synapse.module_api.errors.Codes can be returned by this callback. Returning a boolean is now deprecated.
synapse.module_api.NOT_SPAM, to allow the operation. Other callbacks may still
decide to reject it.
synapse.module_api.errors.Codes to reject the operation with an error code. In case
of doubt, synapse.module_api.errors.Codes.FORBIDDEN is a good error code.
(deprecated) False, which is the same as returning synapse.module_api.NOT_SPAM.
(deprecated) True, which is the same as returning synapse.module_api.errors.Codes.FORBIDDEN.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Called when checking whether a remote server can federate an event with us. Returning
True from this function will silently drop a federated event and split-brain our view
of a room's DAG, and thus you shouldn't use this callback unless you know what you are
doing.
If multiple modules implement this callback, they will be considered in order. If a
callback returns False, Synapse falls through to the next one. The value of the first
callback that does not return False will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
device_id: The device ID the user is re-logging into.
initial_display_name: The device display name, if any.
request_info: A collection of tuples, which first item is a user agent, and which
second item is an IP address. These user agents and IP addresses are the ones that were
used during the login process.
auth_provider_id: The identifier of the SSO authentication provider, if any.
If multiple modules implement this callback, they will be considered in order. If a
callback returns synapse.module_api.NOT_SPAM, Synapse falls through to the next one.
The value of the first callback that does not return synapse.module_api.NOT_SPAM will
be used. If this happens, Synapse will not call any of the subsequent implementations of
this callback.
Note: This will not be called when a user registers.
The example below is a module that implements the spam checker callback
check_event_for_spam to deny any message sent by users whose Matrix user IDs are
mentioned in a configured list, and registers a web resource to the path
/_synapse/client/list_spam_checker/is_evil that returns a JSON object indicating
whether the provided user appears in that list.
import json
from typing import Union
from twisted.web.resource import Resource
from twisted.web.server import Request
from synapse.module_api import ModuleApi
class IsUserEvilResource(Resource):
def __init__(self, config):
super(IsUserEvilResource, self).__init__()
self.evil_users = config.get("evil_users") or []
def render_GET(self, request: Request):
user = request.args.get(b"user")[0].decode()
request.setHeader(b"Content-Type", b"application/json")
return json.dumps({"evil": user in self.evil_users}).encode()
class ListSpamChecker:
def __init__(self, config: dict, api: ModuleApi):
self.api = api
self.evil_users = config.get("evil_users") or []
self.api.register_spam_checker_callbacks(
check_event_for_spam=self.check_event_for_spam,
)
self.api.register_web_resource(
path="/_synapse/client/list_spam_checker/is_evil",
resource=IsUserEvilResource(config),
)
async def check_event_for_spam(self, event: "synapse.events.EventBase") -> Union[Literal["NOT_SPAM"], Codes]:
if event.sender in self.evil_users:
return Codes.FORBIDDEN
else:
return synapse.module_api.NOT_SPAM
Third party rules callbacks allow module developers to add extra checks to verify the
validity of incoming events. Third party event rules callbacks can be registered using
the module API's register_third_party_rules_callbacks method.
This callback is very experimental and can and will break without notice. Module developers
are encouraged to implement check_event_for_spam from the spam checker category instead.
Called when processing any incoming event, with the event and a StateMap
representing the current state of the room the event is being sent into. A StateMap is
a dictionary that maps tuples containing an event type and a state key to the
corresponding state event. For example retrieving the room's m.room.create event from
the state_events argument would look like this: state_events.get(("m.room.create", "")).
The module must return a boolean indicating whether the event can be allowed.
Note that this callback function processes incoming events coming via federation
traffic (on top of client traffic). This means denying an event might cause the local
copy of the room's history to diverge from that of remote servers. This may cause
federation issues in the room. It is strongly recommended to only deny events using this
callback function if the sender is a local user, or in a private federation in which all
servers are using the same module, with the same configuration.
If the boolean returned by the module is True, it may also tell Synapse to replace the
event with new data by returning the new event's data as a dictionary. In order to do
that, it is recommended the module calls event.get_dict() to get the current event as a
dictionary, and modify the returned dictionary accordingly.
If check_event_allowed raises an exception, the module is assumed to have failed.
The event will not be accepted but is not treated as explicitly rejected, either.
An HTTP request causing the module check will likely result in a 500 Internal
Server Error.
When the boolean returned by the module is False, the event is rejected.
(Module developers should not use exceptions for rejection.)
Note that replacing the event only works for events sent by local users, not for events
received over federation.
If multiple modules implement this callback, they will be considered in order. If a
callback returns True, Synapse falls through to the next one. The value of the first
callback that does not return True will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called when processing a room creation request, with the Requester object for the user
performing the request, a dictionary representing the room creation request's JSON body
(see the spec
for a list of possible parameters), and a boolean indicating whether the user performing
the request is a server admin.
Modules can modify the request_content (by e.g. adding events to its initial_state),
or deny the room's creation by raising a module_api.errors.SynapseError.
If multiple modules implement this callback, they will be considered in order. If a
callback returns without raising an exception, Synapse falls through to the next one. The
room creation will be forbidden as soon as one of the callbacks raises an exception. If
this happens, Synapse will not call any of the subsequent implementations of this
callback.
Called when processing an invite via a third-party identifier (i.e. email or phone number).
The module must return a boolean indicating whether the invite can go through.
If multiple modules implement this callback, they will be considered in order. If a
callback returns True, Synapse falls through to the next one. The value of the first
callback that does not return True will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called when changing the visibility of a room in the local public room directory. The
visibility is a string that's either "public" or "private". The module must return a
boolean indicating whether the change can go through.
If multiple modules implement this callback, they will be considered in order. If a
callback returns True, Synapse falls through to the next one. The value of the first
callback that does not return True will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called after sending an event into a room. The module is passed the event, as well
as the state of the room after the event. This means that if the event is a state event,
it will be included in this state.
Note that this callback is called when the event has already been processed and stored
into the room, which means this callback cannot be used to deny persisting the event. To
deny an incoming event, see check_event_for_spam instead.
For any given event, this callback will be called on every worker process, even if that worker will not end up
acting on that event. This callback will not be called for events that are marked as rejected.
If multiple modules implement this callback, Synapse runs them all in order.
Called when an admin user requests the shutdown of a room. The module must return a
boolean indicating whether the shutdown can go through. If the callback returns False,
the shutdown will not proceed and the caller will see a M_FORBIDDEN error.
If multiple modules implement this callback, they will be considered in order. If a
callback returns True, Synapse falls through to the next one. The value of the first
callback that does not return True will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called when the deactivation of a user is requested. User deactivation can be
performed by an admin or the user themselves, so developers are encouraged to check the
requester when implementing this callback. The module must return a
boolean indicating whether the deactivation can go through. If the callback returns False,
the deactivation will not proceed and the caller will see a M_FORBIDDEN error.
The module is passed two parameters, user_id which is the ID of the user being deactivated, and by_admin which is True if the request is made by a serve admin, and False otherwise.
If multiple modules implement this callback, they will be considered in order. If a
callback returns True, Synapse falls through to the next one. The value of the first
callback that does not return True will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called after updating a local user's profile. The update can be triggered either by the
user themselves or a server admin. The update can also be triggered by a user being
deactivated (in which case their display name is set to an empty string ("") and the
avatar URL is set to None). The module is passed the Matrix ID of the user whose profile
has been updated, their new profile, as well as a by_admin boolean that is True if the
update was triggered by a server admin (and False otherwise), and a deactivated
boolean that is True if the update is a result of the user being deactivated.
Note that the by_admin boolean is also True if the profile change happens as a result
of the user logging in through Single Sign-On, or if a server admin updates their own
profile.
Per-room profile changes do not trigger this callback to be called. Synapse administrators
wishing this callback to be called on every profile change are encouraged to disable
per-room profiles globally using the allow_per_room_profiles configuration setting in
Synapse's configuration file.
This callback is not called when registering a user, even when setting it through the
get_displayname_for_registration
module callback.
If multiple modules implement this callback, Synapse runs them all in order.
Called after deactivating a local user, or reactivating them through the admin API. The
deactivation can be triggered either by the user themselves or a server admin. The module
is passed the Matrix ID of the user whose status is changed, as well as a deactivated
boolean that is True if the user is being deactivated and False if they're being
reactivated, and a by_admin boolean that is True if the deactivation was triggered by
a server admin (and False otherwise). This latter by_admin boolean is always True
if the user is being reactivated, as this operation can only be performed through the
admin API.
If multiple modules implement this callback, Synapse runs them all in order.
This callback is deprecated in favour of the on_add_user_third_party_identifier callback, which
features the same functionality. The only difference is in name.
Called after creating an association between a local user and a third-party identifier
(email address, phone number). The module is given the Matrix ID of the user the
association is for, as well as the medium (email or msisdn) and address of the
third-party identifier.
Note that this callback is not called after a successful association on an identity
server.
If multiple modules implement this callback, Synapse runs them all in order.
Called after successfully creating an association between a user and a third-party identifier
(email address, phone number). The module is given the Matrix ID of the user the
association is for, as well as the medium (email or msisdn) and address of the
third-party identifier (i.e. an email address).
Note that this callback is not called if a user attempts to bind their third-party identifier
to an identity server (via a call to POST /_matrix/client/v3/account/3pid/bind).
If multiple modules implement this callback, Synapse runs them all in order.
Called after successfully removing an association between a user and a third-party identifier
(email address, phone number). The module is given the Matrix ID of the user the
association is for, as well as the medium (email or msisdn) and address of the
third-party identifier (i.e. an email address).
Note that this callback is not called if a user attempts to unbind their third-party
identifier from an identity server (via a call to POST /_matrix/client/v3/account/3pid/unbind).
If multiple modules implement this callback, Synapse runs them all in order.
The example below is a module that implements the third-party rules callback
check_event_allowed to censor incoming messages as dictated by a third-party service.
Presence router callbacks allow module developers to define additional users
which receive presence updates from local users. The additional users
can be local or remote.
For example, it could be used to direct all of @alice:example.com (a local user)'s
presence updates to @bob:matrix.org (a remote user), even though they don't share a
room. (Note that those presence updates might not make it to @bob:matrix.org's client
unless a similar presence router is running on that homeserver.)
Presence router callbacks can be registered using the module API's
register_presence_router_callbacks method.
Requiresget_interested_users to also be registered
Called when processing updates to the presence state of one or more users. This callback can
be used to instruct the server to forward that presence state to specific users. The module
must return a dictionary that maps from Matrix user IDs (which can be local or remote) to the
UserPresenceState changes that they should be forwarded.
Synapse will then attempt to send the specified presence updates to each user when possible.
If multiple modules implement this callback, Synapse merges all the dictionaries returned
by the callbacks. If multiple callbacks return a dictionary containing the same key,
Synapse concatenates the sets associated with this key from each dictionary.
Requiresget_users_for_states to also be registered
Called when determining which users someone should be able to see the presence state of. This
callback should return complementary results to get_users_for_state or the presence information
may not be properly forwarded.
The callback is given the Matrix user ID for a local user that is requesting presence data and
should return the Matrix user IDs of the users whose presence state they are allowed to
query. The returned users can be local or remote.
Alternatively the callback can return synapse.module_api.PRESENCE_ALL_USERS
to indicate that the user should receive updates from all known users.
If multiple modules implement this callback, they will be considered in order. Synapse
calls each callback one by one, and use a concatenation of all the sets returned by the
callbacks. If one callback returns synapse.module_api.PRESENCE_ALL_USERS, Synapse uses
this value instead. If this happens, Synapse does not call any of the subsequent
implementations of this callback.
The example below is a module that implements both presence router callbacks, and ensures
that @alice:example.org receives all presence updates from @bob:example.com and
@charlie:somewhere.org, regardless of whether Alice shares a room with any of them.
from typing import Dict, Iterable, Set, Union
from synapse.module_api import ModuleApi
class CustomPresenceRouter:
def __init__(self, config: dict, api: ModuleApi):
self.api = api
self.api.register_presence_router_callbacks(
get_users_for_states=self.get_users_for_states,
get_interested_users=self.get_interested_users,
)
async def get_users_for_states(
self,
state_updates: Iterable["synapse.api.UserPresenceState"],
) -> Dict[str, Set["synapse.api.UserPresenceState"]]:
res = {}
for update in state_updates:
if (
update.user_id == "@bob:example.com"
or update.user_id == "@charlie:somewhere.org"
):
res.setdefault("@alice:example.com", set()).add(update)
return res
async def get_interested_users(
self,
user_id: str,
) -> Union[Set[str], "synapse.module_api.PRESENCE_ALL_USERS"]:
if user_id == "@alice:example.com":
return {"@bob:example.com", "@charlie:somewhere.org"}
return set()
Account validity callbacks allow module developers to add extra steps to verify the
validity on an account, i.e. see if a user can be granted access to their account on the
Synapse instance. Account validity callbacks can be registered using the module API's
register_account_validity_callbacks method.
Called when processing any authenticated request (except for logout requests). The module
can return a bool to indicate whether the user has expired and should be locked out of
their account, or None if the module wasn't able to figure it out. The user is
represented by their Matrix user ID (e.g. @alice:example.com).
If the module returns True, the current request will be denied with the error code
ORG_MATRIX_EXPIRED_ACCOUNT and the HTTP status code 403. Note that this doesn't
invalidate the user's access token.
If multiple modules implement this callback, they will be considered in order. If a
callback returns None, Synapse falls through to the next one. The value of the first
callback that does not return None will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Called after successfully registering a user, in case the module needs to perform extra
operations to keep track of them. (e.g. add them to a database table). The user is
represented by their Matrix user ID.
If multiple modules implement this callback, Synapse runs them all in order.
Called after successfully login or registration of a user for cases when module needs to perform extra operations after auth.
represented by their Matrix user ID.
If multiple modules implement this callback, Synapse runs them all in order.
Password auth providers offer a way for server administrators to integrate
their Synapse installation with an external authentication system. The callbacks can be
registered by using the Module API's register_password_auth_provider_callbacks method.
A dict mapping from tuples of a login type identifier (such as m.login.password) and a
tuple of field names (such as ("password", "secret_thing")) to authentication checking
callbacks, which should be of the following form:
The login type and field names should be provided by the user in the
request to the /login API. The Matrix specification
defines some types, however user defined ones are also allowed.
The callback is passed the user field provided by the client (which might not be in
@username:server form), the login type, and a dictionary of login secrets passed by
the client.
If the authentication is successful, the module must return the user's Matrix ID (e.g.
@alice:example.com) and optionally a callback to be called with the response to the
/login request. If the module doesn't wish to return a callback, it must return None
instead.
If the authentication is unsuccessful, the module must return None.
Note that the user is not automatically registered, the register_user(..) method of
the module API can be used to lazily create users.
If multiple modules register an auth checker for the same login type but with different
fields, Synapse will refuse to start.
If multiple modules register an auth checker for the same login type with the same fields,
then the callbacks will be executed in order, until one returns a Matrix User ID (and
optionally a callback). In that case, the return value of that callback will be accepted
and subsequent callbacks will not be fired. If every callback returns None, then the
authentication fails.
Called when a user attempts to register or log in with a third party identifier,
such as email. It is passed the medium (eg. email), an address (eg. jdoe@example.com)
and the user's password.
If the authentication is successful, the module must return the user's Matrix ID (e.g.
@alice:example.com) and optionally a callback to be called with the response to the /login request.
If the module doesn't wish to return a callback, it must return None instead.
If the authentication is unsuccessful, the module must return None.
If multiple modules implement this callback, they will be considered in order. If a
callback returns None, Synapse falls through to the next one. The value of the first
callback that does not return None will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback. If every callback returns None,
the authentication is denied.
Called during a logout request for a user. It is passed the qualified user ID, the ID of the
deactivated device (if any: access tokens are occasionally created without an associated
device ID), and the (now deactivated) access token.
Deleting the related pushers is done after calling on_logged_out, so you can rely on them
to still be present.
If multiple modules implement this callback, Synapse runs them all in order.
Called when registering a new user. The module can return a username to set for the user
being registered by returning it as a string, or None if it doesn't wish to force a
username for this user. If a username is returned, it will be used as the local part of a
user's full Matrix ID (e.g. it's alice in @alice:example.com).
This callback is called once User-Interactive Authentication
has been completed by the user. It is not called when registering a user via SSO. It is
passed two dictionaries, which include the information that the user has provided during
the registration process.
The first dictionary contains the results of the User-Interactive Authentication
flow followed by the user. Its keys are the identifiers of every step involved in the flow,
associated with either a boolean value indicating whether the step was correctly completed,
or additional information (e.g. email address, phone number...). A list of most existing
identifiers can be found in the Matrix specification.
Here's an example featuring all currently supported keys:
{
"m.login.dummy": True, # Dummy authentication
"m.login.terms": True, # User has accepted the terms of service for the homeserver
"m.login.recaptcha": True, # User has completed the recaptcha challenge
"m.login.email.identity": { # User has provided and verified an email address
"medium": "email",
"address": "alice@example.com",
"validated_at": 1642701357084,
},
"m.login.msisdn": { # User has provided and verified a phone number
"medium": "msisdn",
"address": "33123456789",
"validated_at": 1642701357084,
},
"m.login.registration_token": "sometoken", # User has registered through a registration token
}
The second dictionary contains the parameters provided by the user's client in the request
to /_matrix/client/v3/register. See the Matrix specification
for a complete list of these parameters.
If the module cannot, or does not wish to, generate a username for this user, it must
return None.
If multiple modules implement this callback, they will be considered in order. If a
callback returns None, Synapse falls through to the next one. The value of the first
callback that does not return None will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback. If every callback returns None,
the username provided by the user is used, if any (otherwise one is automatically
generated).
Called when registering a new user. The module can return a display name to set for the
user being registered by returning it as a string, or None if it doesn't wish to force a
display name for this user.
This callback is called once User-Interactive Authentication
has been completed by the user. It is not called when registering a user via SSO. It is
passed two dictionaries, which include the information that the user has provided during
the registration process. These dictionaries are identical to the ones passed to
get_username_for_registration, so refer to the
documentation of this callback for more information about them.
If multiple modules implement this callback, they will be considered in order. If a
callback returns None, Synapse falls through to the next one. The value of the first
callback that does not return None will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback. If every callback returns None,
the username will be used (e.g. alice if the user being registered is @alice:example.com).
Called when attempting to bind a third-party identifier (i.e. an email address or a phone
number). The module is given the medium of the third-party identifier (which is email if
the identifier is an email address, or msisdn if the identifier is a phone number) and
its address, as well as a boolean indicating whether the attempt to bind is happening as
part of registering a new user. The module must return a boolean indicating whether the
identifier can be allowed to be bound to an account on the local homeserver.
If multiple modules implement this callback, they will be considered in order. If a
callback returns True, Synapse falls through to the next one. The value of the first
callback that does not return True will be used. If this happens, Synapse will not call
any of the subsequent implementations of this callback.
Background update controller callbacks allow module developers to control (e.g. rate-limit)
how database background updates are run. A database background update is an operation
Synapse runs on its database in the background after it starts. It's usually used to run
database operations that would take too long if they were run at the same time as schema
updates (which are run on startup) and delay Synapse's startup too much: populating a
table with a big amount of data, adding an index on a big table, deleting superfluous data,
etc.
Background update controller callbacks can be registered using the module API's
register_background_update_controller_callbacks method. Only the first module (in order
of appearance in Synapse's configuration file) calling this method can register background
update controller callbacks, subsequent calls are ignored.
The available background update controller callbacks are:
Called when about to do an iteration of a background update. The module is given the name
of the update, the name of the database, and a flag to indicate whether the background
update will happen in one go and may take a long time (e.g. creating indices). If this last
argument is set to False, the update will be run in batches.
The module must return an async context manager. It will be entered before Synapse runs a
background update; this should return the desired duration of the iteration, in
milliseconds.
The context manager will be exited when the iteration completes. Note that the duration
returned by the context manager is a target, and an iteration may take substantially longer
or shorter. If the one_shot flag is set to True, the duration returned is ignored.
Note: Unlike most module callbacks in Synapse, this one is synchronous. This is
because asynchronous operations are expected to be run by the async context manager.
This callback is required when registering any other background update controller callback.
async def default_batch_size(update_name: str, database_name: str) -> int
Called before the first iteration of a background update, with the name of the update and
of the database. The module must return the number of elements to process in this first
iteration.
If this callback is not defined, Synapse will use a default value of 100.
async def min_batch_size(update_name: str, database_name: str) -> int
Called before running a new batch for a background update, with the name of the update and
of the database. The module must return an integer representing the minimum number of
elements to process in this iteration. This number must be at least 1, and is used to
ensure that progress is always made.
If this callback is not defined, Synapse will use a default value of 100.
Account data callbacks allow module developers to react to changes of the account data
of local users. Account data callbacks can be registered using the module API's
register_account_data_callbacks method.
Called after user's account data has been updated. The module is given the
Matrix ID of the user whose account data is changing, the room ID the data is associated
with, the type associated with the change, as well as the new content. If the account
data is not associated with a specific room, then the room ID is None.
This callback is triggered when new account data is added or when the data associated with
a given type (and optionally room) changes. This includes deletion, since in Matrix,
deleting account data consists of replacing the data associated with a given type
(and optionally room) with an empty dictionary ({}).
Note that this doesn't trigger when changing the tags associated with a room, as these are
processed separately by Synapse.
If multiple modules implement this callback, Synapse runs them all in order.
The example below is a module that implements the on_account_data_updated callback, and
sends an event to an audit room when a user changes their account data.
import json
import attr
from typing import Any, Dict, Optional
from synapse.module_api import JsonDict, ModuleApi
from synapse.module_api.errors import ConfigError
@attr.s(auto_attribs=True)
class CustomAccountDataConfig:
audit_room: str
sender: str
class CustomAccountDataModule:
def __init__(self, config: CustomAccountDataConfig, api: ModuleApi):
self.api = api
self.config = config
self.api.register_account_data_callbacks(
on_account_data_updated=self.log_new_account_data,
)
@staticmethod
def parse_config(config: Dict[str, Any]) -> CustomAccountDataConfig:
def check_in_config(param: str):
if param not in config:
raise ConfigError(f"'{param}' is required")
check_in_config("audit_room")
check_in_config("sender")
return CustomAccountDataConfig(
audit_room=config["audit_room"],
sender=config["sender"],
)
async def log_new_account_data(
self,
user_id: str,
room_id: Optional[str],
account_data_type: str,
content: JsonDict,
) -> None:
content_raw = json.dumps(content)
msg_content = f"{user_id} has changed their account data for type {account_data_type} to: {content_raw}"
if room_id is not None:
msg_content += f" (in room {room_id})"
await self.api.create_and_send_event_into_room(
{
"room_id": self.config.audit_room,
"sender": self.config.sender,
"type": "m.room.message",
"content": {
"msgtype": "m.text",
"body": msg_content
}
}
)
In order to port a module that uses Synapse's old module interface, its author needs to:
ensure the module's callbacks are all asynchronous.
register their callbacks using one or more of the register_[...]_callbacks methods
from the ModuleApi class in the module's __init__ method (see this section
for more info).
Additionally, if the module is packaged with an additional web resource, the module
should register this resource in its __init__ method using the register_web_resource
method from the ModuleApi class (see this section for
more info).
There is no longer a get_db_schema_files callback provided for password auth provider modules. Any
changes to the database should now be made by the module using the module API class.
The module's author should also update any example in the module's configuration to only
use the new modules section in Synapse's configuration file (see this section
for more info).
For small instances it is recommended to run Synapse in the default monolith mode.
For larger instances where performance is a concern it can be helpful to split
out functionality into multiple separate python processes. These processes are
called 'workers', and are (eventually) intended to scale horizontally
independently.
Synapse's worker support is under active development and subject to change as
we attempt to rapidly scale ever larger Synapse instances. However we are
documenting it here to help admins needing a highly scalable Synapse instance
similar to the one running matrix.org.
All processes continue to share the same database instance, and as such,
workers only work with PostgreSQL-based Synapse deployments. SQLite should only
be used for demo purposes and any admin considering workers should already be
running PostgreSQL.
The processes communicate with each other via a Synapse-specific protocol called
'replication' (analogous to MySQL- or Postgres-style database replication) which
feeds streams of newly written data between processes so they can be kept in
sync with the database state.
When configured to do so, Synapse uses a
Redis pub/sub channel to send the replication
stream between all configured Synapse processes. Additionally, processes may
make HTTP requests to each other, primarily for operations which need to wait
for a reply ─ such as sending an event.
All the workers and the main process connect to Redis, which relays replication
commands between processes.
If Redis support is enabled Synapse will use it as a shared cache, as well as a
pub/sub mechanism.
See the Architectural diagram section at the end for
a visualisation of what this looks like.
A Redis server is required to manage the communication between the processes.
The Redis server should be installed following the normal procedure for your
distribution (e.g. apt install redis-server on Debian). It is safe to use an
existing Redis deployment if you have one.
Once installed, check that Redis is running and accessible from the host running
Synapse, for example by executing echo PING | nc -q1 localhost 6379 and seeing
a response of +PONG.
The appropriate dependencies must also be installed for Synapse. If using a
virtualenv, these can be installed with:
pip install "matrix-synapse[redis]"
Note that these dependencies are included when synapse is installed with pip install matrix-synapse[all]. They are also included in the debian packages from
matrix.org and in the docker images at
https://hub.docker.com/r/matrixdotorg/synapse/.
To make effective use of the workers, you will need to configure an HTTP
reverse-proxy such as nginx or haproxy, which will direct incoming requests to
the correct worker, or to the main synapse instance. See
the reverse proxy documentation for information on setting up a reverse
proxy.
When using workers, each worker process has its own configuration file which
contains settings specific to that worker, such as the HTTP listener that it
provides (if any), logging configuration, etc.
Normally, the worker processes are configured to read from a shared
configuration file as well as the worker-specific configuration files. This
makes it easier to keep common configuration settings synchronised across all
the processes.
The main process is somewhat special in this respect: it does not normally
need its own configuration file and can take all of its configuration from the
shared configuration file.
You will need to add an instance_map
with the main process defined, as well as the relevant connection information from
it's HTTP replication listener (defined in step 1 above).
Note that the host defined is the address the worker needs to look for the main
process at, not necessarily the same address that is bound to.
If you are using Unix sockets for the replication resource, make sure to
use a path to the socket file instead of a port.
Optionally, a shared secret
can be used to authenticate HTTP traffic between workers. For example:
# extend the existing `listeners` section. This defines the ports that the
# main process will listen on.
listeners:
# The HTTP replication port
- port: 9093
bind_address: '127.0.0.1'
type: http
resources:
- names: [replication]
# Add a random shared secret to authenticate traffic.
worker_replication_secret: ""
redis:
enabled: true
instance_map:
main:
host: 'localhost'
port: 9093
If handling HTTP requests, a worker_listeners option
with an http listener.
Synapse 1.72 and older: if handling the ^/_matrix/client/v3/keys/upload endpoint, the HTTP URI for
the main process (worker_main_http_uri). This config option is no longer required and is ignored when running Synapse 1.73 and newer.
...is a full configuration for a generic worker instance, which will expose a
plain HTTP endpoint on port 8083 separately serving various endpoints, e.g.
/sync, which are listed below.
Obviously you should configure your reverse-proxy to route the relevant
endpoints to the worker (localhost:8083 in the above example).
Finally, you need to start your worker processes. This can be done with either
synctl or your distribution's preferred service manager such as systemd. We
recommend the use of systemd where available: for information on setting up
systemd to start synapse workers, see
Systemd with Workers. To use synctl, see
Using synctl with Workers.
This worker can handle API requests matching the following regular expressions.
These endpoints can be routed to any worker. If a worker is set up to handle a
stream then, for maximum efficiency, additional endpoints should be routed to that
worker: refer to the stream writers section below for further
information.
Pagination requests can also be handled, but all requests for a given
room must be routed to the same instance. Additionally, care must be taken to
ensure that the purge history admin API is not used while pagination requests
for the room are in flight:
Additionally, the following endpoints should be included if Synapse is configured
to use SSO (you only need to include the ones for whichever SSO provider you're
using):
# for all SSO providers
^/_matrix/client/(api/v1|r0|v3|unstable)/login/sso/redirect
^/_synapse/client/pick_idp$
^/_synapse/client/pick_username
^/_synapse/client/new_user_consent$
^/_synapse/client/sso_register$
# OpenID Connect requests.
^/_synapse/client/oidc/callback$
# SAML requests.
^/_synapse/client/saml2/authn_response$
# CAS requests.
^/_matrix/client/(api/v1|r0|v3|unstable)/login/cas/ticket$
Ensure that all SSO logins go to a single process.
For multiple workers not handling the SSO endpoints properly, see
#7530 and
#9427.
Note that a HTTP listener
with client and federationresources must be configured in the
worker_listeners
option in the worker config.
It is possible to run multiple instances of this worker app, with incoming requests
being load-balanced between them by the reverse-proxy. However, different endpoints
have different characteristics and so admins
may wish to run multiple groups of workers handling different endpoints so that
load balancing can be done in different ways.
For /sync and /initialSync requests it will be more efficient if all
requests from a particular user are routed to a single instance. This can
be done in reverse proxy by extracting username part from the users access token.
Admins may additionally wish to separate out /sync
requests that have a since query parameter from those that don't (and
/initialSync), as requests that don't are known as "initial sync" that happens
when a user logs in on a new device and can be very resource intensive, so
isolating these requests will stop them from interfering with other users ongoing
syncs.
Example nginx configuration snippet that handles the cases above. This is just an
example and probably requires some changes according to your particular setup:
# Choose sync worker based on the existence of "since" query parameter
map $arg_since $sync {
default synapse_sync;
'' synapse_initial_sync;
}
# Extract username from access token passed as URL parameter
map $arg_access_token $accesstoken_from_urlparam {
# Defaults to just passing back the whole accesstoken
default $arg_access_token;
# Try to extract username part from accesstoken URL parameter
"~syt_(?<username>.*?)_.*" $username;
}
# Extract username from access token passed as authorization header
map $http_authorization $mxid_localpart {
# Defaults to just passing back the whole accesstoken
default $http_authorization;
# Try to extract username part from accesstoken header
"~Bearer syt_(?<username>.*?)_.*" $username;
# if no authorization-header exist, try mapper for URL parameter "access_token"
"" $accesstoken_from_urlparam;
}
upstream synapse_initial_sync {
# Use the username mapper result for hash key
hash $mxid_localpart consistent;
server 127.0.0.1:8016;
server 127.0.0.1:8036;
}
upstream synapse_sync {
# Use the username mapper result for hash key
hash $mxid_localpart consistent;
server 127.0.0.1:8013;
server 127.0.0.1:8037;
server 127.0.0.1:8038;
server 127.0.0.1:8039;
}
# Sync initial/normal
location ~ ^/_matrix/client/(r0|v3)/sync$ {
proxy_pass http://$sync;
}
# Normal sync
location ~ ^/_matrix/client/(api/v1|r0|v3)/events$ {
proxy_pass http://synapse_sync;
}
# Initial_sync
location ~ ^/_matrix/client/(api/v1|r0|v3)/initialSync$ {
proxy_pass http://synapse_initial_sync;
}
location ~ ^/_matrix/client/(api/v1|r0|v3)/rooms/[^/]+/initialSync$ {
proxy_pass http://synapse_initial_sync;
}
Federation and client requests can be balanced via simple round robin.
The inbound federation transaction request ^/_matrix/federation/v1/send/
should be balanced by source IP so that transactions from the same remote server
go to the same process.
Registration/login requests can be handled separately purely to help ensure that
unexpected load doesn't affect new logins and sign ups.
Finally, event sending requests can be balanced by the room ID in the URI (or
the full URI, or even just round robin), the room ID is the path component after
/rooms/. If there is a large bridge connected that is sending or may send lots
of events, then a dedicated set of workers can be provisioned to limit the
effects of bursts of events from that bridge on events sent by normal users.
worker_app: synapse.app.generic_worker
worker_name: event_persister1
worker_listeners:
- type: http
port: 8034
resources:
- names: [replication]
# Enable listener if this stream writer handles endpoints for the `typing` or
# `to_device` streams. Uses a different port to the `replication` listener to
# avoid exposing the `replication` listener publicly.
#
#- type: http
# port: 8035
# x_forwarded: true
# resources:
# - names: [client]
worker_log_config: /etc/matrix-synapse/event-persister-log.yaml
Some of the streams have associated endpoints which, for maximum efficiency, should
be routed to the workers handling that stream. See below for the currently supported
streams and the endpoints associated with them:
The events stream experimentally supports having multiple writer workers, where load
is sharded between them by room ID. Each writer is called an event persister. They are
responsible for
Because load is sharded in this way, you must restart all worker instances when
adding or removing event persisters.
An event_persister should not be mistaken for an event_creator.
An event_creator listens for requests from clients to create new events and does
so. It will then pass those events over HTTP replication to any configured event
persisters (or the main process if none are configured).
Note that event_creators and event_persisters are implemented using the same
synapse.app.generic_worker.
An example stream_writers
configuration with multiple writers:
The
outbound_federation_restricted_to
configuration is useful to make sure outbound federation traffic only goes through a
specified subset of workers. This allows you to set more strict access controls (like a
firewall) for all workers and only allow the federation_sender's to contact the
outside world.
There is also support for moving background tasks to a separate
worker. Background tasks are run periodically or started via replication. Exactly
which tasks are configured to run depends on your Synapse configuration (e.g. if
stats is enabled). This worker doesn't handle any REST endpoints itself.
To enable this, the worker must have a unique
worker_name
and can be configured to run background tasks. For example, to move background tasks
to a dedicated worker, the shared configuration would include:
You can designate one generic worker to send output traffic to Application Services.
Doesn't handle any REST endpoints itself, but you should specify its name in the
shared configuration
as follows:
notify_appservices_from_worker: worker_name
This work cannot be load-balanced; please ensure the main process is restarted
after setting this option in the shared configuration!
This style of configuration supersedes the legacy synapse.app.appservice
worker application type.
Multiple workers can be added to this map, in which case the work is balanced
across them. Ensure the main process and all pusher workers are restarted after changing
this option.
These workers don't need to accept incoming HTTP requests to send push notifications,
so no additional reverse proxy configuration is required for pusher workers.
This style of configuration supersedes the legacy synapse.app.pusher
worker application type.
Handles sending push notifications to sygnal and email. Doesn't handle any
REST endpoints itself, but you should set
start_pushers: false in the
shared configuration file to stop the main synapse sending push notifications.
To run multiple instances at once the
pusher_instances
option should list all pusher instances by their
worker_name, e.g.:
Handles sending output traffic to Application Services. Doesn't handle any
REST endpoints itself, but you should set notify_appservices: False in the
shared configuration file to stop the main synapse sending appservice notifications.
Note this worker cannot be load-balanced: only one instance should be active.
Handles sending federation traffic to other servers. Doesn't handle any
REST endpoints itself, but you should set
send_federation: false
in the shared configuration file to stop the main synapse sending this traffic.
If running multiple federation senders then you must list each
instance in the
federation_sender_instances
option by their
worker_name.
All instances must be stopped and started when adding or removing instances.
For example:
You should also set
enable_media_repo: False
in the shared configuration
file to stop the main synapse running background jobs related to managing the
media repository. Note that doing so will prevent the main process from being
able to handle the above endpoints.
In the media_repository worker configuration file, configure the
HTTP listener to
expose the media resource. For example:
Note that if running multiple media repositories they must be on the same server
and you must specify a single instance to run the background tasks in the
shared configuration,
e.g.:
Note that if a reverse proxy is used , then /_matrix/media/ must be routed for both inbound client and federation requests (if they are handled separately).
When using this worker you must also set update_user_directory: false in the
shared configuration file to stop the main synapse running background
jobs related to updating the user directory.
Above endpoint is not required to be routed to this worker. By default,
update_user_directory is set to true, which means the main process
will handle updates. All workers configured with client can handle the above
endpoint as long as either this worker or the main process are configured to
handle it, and are online.
If update_user_directory is set to false, and this worker is not running,
the above endpoint may give outdated results.
A main change that has occurred is the merging of worker apps into
synapse.app.generic_worker. This change is backwards compatible and so no
changes to the config are required.
To migrate apps to use synapse.app.generic_worker simply update the
worker_app option in the worker configs, and where worker are started (e.g.
in systemd service files, but not required for synctl).
If you want to use synctl to manage your synapse processes, you will need to
create an an additional configuration file for the main synapse process. That
configuration should look like this:
worker_app: synapse.app.homeserver
Additionally, each worker app must be configured with the name of a "pid file",
to which it will write its process ID when it starts. For example, for a
synchrotron, you might write:
worker_pid_file: /home/matrix/synapse/worker1.pid
Finally, to actually run your worker-based synapse, you must pass synctl the -a
commandline option to tell it to operate on all the worker configurations found
in the given directory, e.g.:
synctl -a $CONFIG/workers start
Currently one should always restart all workers when restarting or upgrading
synapse, unless you explicitly know it's safe not to. For instance, restarting
synapse without restarting all the synchrotrons may result in broken typing
notifications.
To manipulate a specific worker, you pass the -w option to synctl:
This is a setup for managing synapse with systemd, including support for
managing workers. It provides a matrix-synapse service for the master, as
well as a matrix-synapse-worker@ service template for any workers you
require. Additionally, to group the required services, it sets up a
matrix-synapse.target.
See the worker documentation for information on how to set up the
configuration files and reverse-proxy correctly.
Below is a sample generic_worker worker configuration file.
Systemd manages daemonization itself, so ensure that none of the configuration
files set either daemonize or worker_daemonize.
The config files of all workers are expected to be located in
/etc/matrix-synapse/workers. If you want to use a different location, edit
the provided *.service files accordingly.
There is no need for a separate configuration file for the master process.
Copy the *.service and *.target files in system
to /etc/systemd/system.
Run systemctl daemon-reload to tell systemd to load the new unit files.
Run systemctl enable matrix-synapse.service. This will configure the
synapse master process to be started as part of the matrix-synapse.target
target.
For each worker process to be enabled, run systemctl enable matrix-synapse-worker@<worker_name>.service. For each <worker_name>, there
should be a corresponding configuration file.
/etc/matrix-synapse/workers/<worker_name>.yaml.
Start all the synapse processes with systemctl start matrix-synapse.target.
Tell systemd to start synapse on boot with systemctl enable matrix-synapse.target.
Once the services are correctly set up, you can use the following commands
to manage your synapse installation:
# Restart Synapse master and all workers
systemctl restart matrix-synapse.target
# Stop Synapse and all workers
systemctl stop matrix-synapse.target
# Restart the master alone
systemctl start matrix-synapse.service
# Restart a specific worker (eg. generic_worker); the master is
# unaffected by this.
systemctl restart matrix-synapse-worker@generic_worker.service
# Add a new worker (assuming all configs are set up already)
systemctl enable matrix-synapse-worker@federation_writer.service
systemctl restart matrix-synapse.target
Optional: If further hardening is desired, the file
override-hardened.conf may be copied from
contrib/systemd/override-hardened.conf
in this repository to the location
/etc/systemd/system/matrix-synapse.service.d/override-hardened.conf (the
directory may have to be created). It enables certain sandboxing features in
systemd to further secure the synapse service. You may read the comments to
understand what the override file is doing. The same file will need to be copied to
/etc/systemd/system/matrix-synapse-worker@.service.d/override-hardened-worker.conf
(this directory may also have to be created) in order to apply the same
hardening options to any worker processes.
Once these files have been copied to their appropriate locations, simply reload
systemd's manager config files and restart all Synapse services to apply the hardening options. They will automatically
be applied at every restart as long as the override files are present at the
specified locations.
In order to see their effect, you may run systemd-analyze security matrix-synapse.service before and after applying the hardening options to see
the changes being applied at a glance.
Many of the API calls in the admin api will require an access_token for a
server admin. (Note that a server admin is distinct from a room admin.)
An existing user can be marked as a server admin by updating the database directly.
Check your database settings in the configuration file, connect to the correct database using either psql [database name] (if using PostgreSQL) or sqlite3 path/to/your/database.db (if using SQLite) and elevate the user @foo:bar.com to administrator.
UPDATE users SET admin = 1 WHERE name = '@foo:bar.com';
A new server admin user can also be created using the register_new_matrix_user
command. This is a script that is distributed as part of synapse. It is possibly
already on your $PATH depending on how Synapse was installed.
Finding your user's access_token is client-dependent, but will usually be shown in the client's settings.
For security reasons, we recommend
that the Admin API (/_synapse/admin/...) should be hidden from public view using a
reverse proxy. This means you should typically query the Admin API from a terminal on
the machine which runs Synapse.
Once you have your access_token, you will need to authenticate each request to an Admin API endpoint by
providing the token as either a query parameter or a request header. To add it as a request header in cURL:
For example, suppose we want to
query the account of the user
@foo:bar.com. We need an admin access token (e.g.
syt_AjfVef2_L33JNpafeif_0feKJfeaf0CQpoZk), and we need to know which port
Synapse's client listener is listening
on (e.g. 8008). Then we can use the following command to request the account
information from the Admin API.
curl --header "Authorization: Bearer syt_AjfVef2_L33JNpafeif_0feKJfeaf0CQpoZk" -X GET http://127.0.0.1:8008/_synapse/admin/v2/users/@foo:bar.com
For more details on access tokens in Matrix, please refer to the complete
matrix spec documentation.
Note: This API is disabled when MSC3861 is enabled. See #15582
This API allows a server administrator to manage the validity of an account. To
use it, you must enable the account validity feature (under
account_validity) in Synapse's configuration.
To use it, you will need to authenticate by providing an access_token
for a server admin: see Admin API.
enabled whether the background updates are enabled or disabled.
db_name the database name (usually Synapse is configured with a single database named 'master').
For each update:
name the name of the update.
total_item_count total number of "items" processed (the meaning of 'items' depends on the update in question).
total_duration_ms how long the background process has been running, not including time spent sleeping.
average_items_per_ms how many items are processed per millisecond based on an exponential average.
Background updates should not be paused for significant periods of time, as
this can affect the performance of Synapse.
Note: This won't persist over restarts.
Note: This won't cancel any update query that is currently running. This is
usually fine since most queries are short lived, except for CREATE INDEX
background updates which won't be cancelled once started.
The API is:
POST /_synapse/admin/v1/background_updates/enabled
with the following body:
{
"enabled": false
}
enabled sets whether the background updates are enabled or disabled.
The API returns the enabled param.
{
"enabled": false
}
There is also a GET version which returns the enabled state.
To paginate, check for next_token and if present, call the endpoint again with from
set to the value of next_token. This will return a new page.
If the endpoint does not return a next_token then there are no more reports to
paginate through.
URL parameters:
limit: integer - Is optional but is used for pagination, denoting the maximum number
of items to return in this call. Defaults to 100.
from: integer - Is optional but used for pagination, denoting the offset in the
returned results. This should be treated as an opaque value and not explicitly set to
anything other than the return value of next_token from a previous call. Defaults to 0.
dir: string - Direction of event report order. Whether to fetch the most recent
first (b) or the oldest first (f). Defaults to b.
user_id: string - Is optional and filters to only return users with user IDs that
contain this value. This is the user who reported the event and wrote the reason.
room_id: string - Is optional and filters to only return rooms with room IDs that
contain this value.
Response
The following fields are returned in the JSON response body:
id: integer - ID of event report.
received_ts: integer - The timestamp (in milliseconds since the unix epoch) when this
report was sent.
room_id: string - The ID of the room in which the event being reported is located.
name: string - The name of the room.
event_id: string - The ID of the reported event.
user_id: string - This is the user who reported the event and wrote the reason.
reason: string - Comment made by the user_id in this report. May be blank or null.
score: integer - Content is reported based upon a negative score, where -100 is
"most offensive" and 0 is "inoffensive". May be null.
sender: string - This is the ID of the user who sent the original message/event that
was reported.
canonical_alias: string - The canonical alias of the room. null if the room does not
have a canonical alias set.
next_token: integer - Indication for pagination. See above.
total: integer - Total number of event reports related to the query
(user_id and room_id).
This API allows a server administrator to enable experimental features for a given user. The request must
provide a body containing the user id and listing the features to enable/disable in the following format:
This API quarantines all local and remote media in a room.
Request:
POST /_synapse/admin/v1/room/<room_id>/media/quarantine
{}
Where room_id is in the form of !roomid12345:example.org.
Response:
{
"num_quarantined": 10
}
The following fields are returned in the JSON response body:
num_quarantined: integer - The number of media items successfully quarantined
Note that there is a legacy endpoint, POST /_synapse/admin/v1/quarantine_media/<room_id>, that operates the same.
However, it is deprecated and may be removed in a future release.
This API quarantines all local media that a local user has uploaded. That is to say, if
you would like to quarantine media uploaded by a user on a remote homeserver, you should
instead use one of the other APIs.
Request:
POST /_synapse/admin/v1/user/<user_id>/media/quarantine
{}
URL Parameters
user_id: string - User ID in the form of @bob:example.org
Response:
{
"num_quarantined": 10
}
The following fields are returned in the JSON response body:
num_quarantined: integer - The number of media items successfully quarantined
This API protects a single piece of local media from being quarantined using the
above APIs. This is useful for sticker packs and other shared media which you do
not want to get quarantined, especially when
quarantining media in a room.
Request:
POST /_synapse/admin/v1/media/protect/<media_id>
{}
This API deletes the local media from the disk of your own server.
This includes any local thumbnails and copies of media downloaded from
remote homeservers.
This API will not affect media that has been uploaded to external
media repositories (e.g https://github.com/turt2live/matrix-media-repo/).
See also Purge Remote Media API.
POST /_synapse/admin/v1/media/delete?before_ts=<before_ts>
{}
Deprecated in Synapse v1.78.0: This API is available at the deprecated endpoint:
POST /_synapse/admin/v1/media/<server_name>/delete?before_ts=<before_ts>
{}
URL Parameters
server_name: string - The name of your local server (e.g matrix.org). Deprecated in Synapse v1.78.0.
before_ts: string representing a positive integer - Unix timestamp in milliseconds.
Files that were last used before this timestamp will be deleted. It is the timestamp of
last access, not the timestamp when the file was created.
size_gt: Optional - string representing a positive integer - Size of the media in bytes.
Files that are larger will be deleted. Defaults to 0.
keep_profiles: Optional - string representing a boolean - Switch to also delete files
that are still used in image data (e.g user profile, room avatar).
If false these files will be deleted. Defaults to true.
The purge remote media API allows server admins to purge old cached remote media.
The API is:
POST /_synapse/admin/v1/purge_media_cache?before_ts=<unix_timestamp_in_ms>
{}
URL Parameters
before_ts: string representing a positive integer - Unix timestamp in milliseconds.
All cached media that was last accessed before this timestamp will be removed.
Response:
{
"deleted": 10
}
The following fields are returned in the JSON response body:
deleted: integer - The number of media items successfully deleted
If the user re-requests purged remote media, synapse will re-request the media
from the originating server.
The purge history API allows server admins to purge historic events from their
database, reclaiming disk space.
Depending on the amount of history being purged a call to the API may take
several minutes or longer. During this period users will not be able to
paginate further back in the room from the point being purged from.
Note that Synapse requires at least one message in each room, so it will never
delete the last message in a room.
To use it, you will need to authenticate by providing an access_token
for a server admin: see Admin API.
The API is:
POST /_synapse/admin/v1/purge_history/<room_id>[/<event_id>]
By default, events sent by local users are not deleted, as they may represent
the only copies of this content in existence. (Events sent by remote users are
deleted.)
Room state data (such as joins, leaves, topic) is always preserved.
To delete local message events as well, set delete_local_events in the body:
{
"delete_local_events": true
}
The caller must specify the point in the room to purge up to. This can be
specified by including an event_id in the URI, or by setting a
purge_up_to_event_id or purge_up_to_ts in the request body. If an event
id is given, that event (and others at the same graph depth) will be retained.
If purge_up_to_ts is given, it should be a timestamp since the unix epoch,
in milliseconds.
The API starts the purge running, and returns immediately with a JSON body with
a purge id:
Note: This API is disabled when MSC3861 is enabled. See #15582
This API allows for the creation of users in an administrative and
non-interactive way. This is generally used for bootstrapping a Synapse
instance with administrator accounts.
To authenticate yourself to the server, you will need both the shared secret
(registration_shared_secret
in the homeserver configuration), and a one-time nonce. If the registration
shared secret is not configured, this API is not enabled.
To fetch the nonce, you need to request one from the API:
> GET /_synapse/admin/v1/register
< {"nonce": "thisisanonce"}
Once you have the nonce, you can make a POST to the same URL with a JSON
body containing the nonce, username, password, whether they are an admin
(optional, False by default), and a HMAC digest of the content. Also you can
set the displayname (optional, username by default).
The MAC is the hex digest output of the HMAC-SHA1 algorithm, with the key being
the shared secret and the content being the nonce, user, password, either the
string "admin" or "notadmin", and optionally the user_type
each separated by NULs.
Here is an easy way to generate the HMAC digest if you have Bash and OpenSSL:
# Update these values and then paste this code block into a bash terminal
nonce='thisisanonce'
username='pepper_roni'
password='pizza'
admin='admin'
secret='shared_secret'
printf '%s\0%s\0%s\0%s' "$nonce" "$username" "$password" "$admin" |
openssl sha1 -hmac "$secret" |
awk '{print $2}'
Note: This API is disabled when MSC3861 is enabled. See #15582
This API allows you to manage tokens which can be used to authenticate
registration requests, as proposed in
MSC3231
and stabilised in version 1.2 of the Matrix specification.
To use it, you will need to enable the registration_requires_token config
option, and authenticate by providing an access_token for a server admin:
see Admin API.
Most endpoints make use of JSON objects that contain details about tokens.
These objects have the following fields:
token: The token which can be used to authenticate registration.
uses_allowed: The number of times the token can be used to complete a
registration before it becomes invalid.
pending: The number of pending uses the token has. When someone uses
the token to authenticate themselves, the pending counter is incremented
so that the token is not used more than the permitted number of times.
When the person completes registration the pending counter is decremented,
and the completed counter is incremented.
completed: The number of times the token has been used to successfully
complete a registration.
expiry_time: The latest time the token is valid. Given as the number of
milliseconds since 1970-01-01 00:00:00 UTC (the start of the Unix epoch).
To convert this into a human-readable form you can remove the milliseconds
and use the date command. For example, date -d '@1625394937'.
Lists all tokens and details about them. If the request is successful, the top
level JSON object will have a registration_tokens key which is an array of
registration token objects.
GET /_synapse/admin/v1/registration_tokens
Optional query parameters:
valid: true or false. If true, only valid tokens are returned.
If false, only tokens that have expired or have had all uses exhausted are
returned. If omitted, all tokens are returned regardless of validity.
Create a new registration token. If the request is successful, the newly created
token will be returned as a registration token object in the response body.
POST /_synapse/admin/v1/registration_tokens/new
The request body must be a JSON object and can contain the following fields:
token: The registration token. A string of no more than 64 characters that
consists only of characters matched by the regex [A-Za-z0-9._~-].
Default: randomly generated.
uses_allowed: The integer number of times the token can be used to complete
a registration before it becomes invalid.
Default: null (unlimited uses).
expiry_time: The latest time the token is valid. Given as the number of
milliseconds since 1970-01-01 00:00:00 UTC (the start of the Unix epoch).
You could use, for example, date '+%s000' -d 'tomorrow'.
Default: null (token does not expire).
length: The length of the token randomly generated if token is not
specified. Must be between 1 and 64 inclusive. Default: 16.
If a field is omitted the default is used.
Example using defaults:
POST /_synapse/admin/v1/registration_tokens/new
{}
Update the number of allowed uses or expiry time of a token. If the request is
successful, the updated token will be returned as a registration token object
in the response body.
PUT /_synapse/admin/v1/registration_tokens/<token>
Path parameters:
token: The registration token to update.
The request body must be a JSON object and can contain the following fields:
uses_allowed: The integer number of times the token can be used to complete
a registration before it becomes invalid. By setting uses_allowed to 0
the token can be easily made invalid without deleting it.
If null the token will have an unlimited number of uses.
expiry_time: The latest time the token is valid. Given as the number of
milliseconds since 1970-01-01 00:00:00 UTC (the start of the Unix epoch).
If null the token will not expire.
If a field is omitted its value is not modified.
Example:
PUT /_synapse/admin/v1/registration_tokens/defg
{
"expiry_time": 4781243146000 // 2121-07-06 11:05:46 UTC
}
This API allows an administrator to join an user account with a given user_id
to a room with a given room_id_or_alias. You can only modify the membership of
local users. The server administrator must be in the room and have permission to
invite users.
To use it, you will need to authenticate by providing an access_token
for a server admin: see Admin API.
The List Room admin API allows server admins to get a list of rooms on their
server. There are various parameters available that allow for filtering and
sorting the returned list. This API supports pagination.
To use it, you will need to authenticate by providing an access_token
for a server admin: see Admin API.
Parameters
The following query parameters are available:
from - Offset in the returned list. Defaults to 0.
limit - Maximum amount of rooms to return. Defaults to 100.
order_by - The method in which to sort the returned list of rooms. Valid values are:
alphabetical - Same as name. This is deprecated.
size - Same as joined_members. This is deprecated.
name - Rooms are ordered alphabetically by room name. This is the default.
canonical_alias - Rooms are ordered alphabetically by main alias address of the room.
joined_members - Rooms are ordered by the number of members. Largest to smallest.
joined_local_members - Rooms are ordered by the number of local members. Largest to smallest.
version - Rooms are ordered by room version. Largest to smallest.
creator - Rooms are ordered alphabetically by creator of the room.
encryption - Rooms are ordered alphabetically by the end-to-end encryption algorithm.
federatable - Rooms are ordered by whether the room is federatable.
public - Rooms are ordered by visibility in room list.
join_rules - Rooms are ordered alphabetically by join rules of the room.
guest_access - Rooms are ordered alphabetically by guest access option of the room.
history_visibility - Rooms are ordered alphabetically by visibility of history of the room.
state_events - Rooms are ordered by number of state events. Largest to smallest.
dir - Direction of room order. Either f for forwards or b for backwards. Setting
this value to b will reverse the above sort order. Defaults to f.
search_term - Filter rooms by their room name, canonical alias and room id.
Specifically, rooms are selected if the search term is contained in
the room's name,
the local part of the room's canonical alias, or
the complete (local and server part) room's id (case sensitive).
Defaults to no filtering.
Response
The following fields are possible in the JSON response body:
rooms - An array of objects, each containing information about a room.
Room objects contain the following fields:
room_id - The ID of the room.
name - The name of the room.
canonical_alias - The canonical (main) alias address of the room.
joined_members - How many users are currently in the room.
joined_local_members - How many local users are currently in the room.
version - The version of the room as a string.
creator - The user_id of the room creator.
encryption - Algorithm of end-to-end encryption of messages. Is null if encryption is not active.
federatable - Whether users on other servers can join this room.
public - Whether the room is visible in room directory.
join_rules - The type of rules used for users wishing to join this room. One of: ["public", "knock", "invite", "private"].
guest_access - Whether guests can join the room. One of: ["can_join", "forbidden"].
history_visibility - Who can see the room history. One of: ["invited", "joined", "shared", "world_readable"].
state_events - Total number of state_events of a room. Complexity of the room.
room_type - The type of the room taken from the room's creation event; for example "m.space" if the room is a space. If the room does not define a type, the value will be null.
offset - The current pagination offset in rooms. This parameter should be
used instead of next_token for room offset as next_token is
not intended to be parsed.
total_rooms - The total number of rooms this query can return. Using this
and offset, you have enough information to know the current
progression through the list.
next_batch - If this field is present, we know that there are potentially
more rooms on the server that did not all fit into this response.
We can use next_batch to get the "next page" of results. To do
so, simply repeat your request, setting the from parameter to
the value of next_batch.
prev_batch - If this field is present, it is possible to paginate backwards.
Use prev_batch for the from value in the next request to
get the "previous page" of results.
The presence of the next_token parameter tells us that there are more rooms
than returned in this request, and we need to make another request to get them.
To get the next batch of room results, we repeat our request, setting the from
parameter to the value of next_token.
GET /_synapse/admin/v1/rooms?order_by=size&from=100
The Room Details admin API allows server admins to get all details of a room.
The following fields are possible in the JSON response body:
room_id - The ID of the room.
name - The name of the room.
topic - The topic of the room.
avatar - The mxc URI to the avatar of the room.
canonical_alias - The canonical (main) alias address of the room.
joined_members - How many users are currently in the room.
joined_local_members - How many local users are currently in the room.
joined_local_devices - How many local devices are currently in the room.
version - The version of the room as a string.
creator - The user_id of the room creator.
encryption - Algorithm of end-to-end encryption of messages. Is null if encryption is not active.
federatable - Whether users on other servers can join this room.
public - Whether the room is visible in room directory.
join_rules - The type of rules used for users wishing to join this room. One of: ["public", "knock", "invite", "private"].
guest_access - Whether guests can join the room. One of: ["can_join", "forbidden"].
history_visibility - Who can see the room history. One of: ["invited", "joined", "shared", "world_readable"].
state_events - Total number of state_events of a room. Complexity of the room.
room_type - The type of the room taken from the room's creation event; for example "m.space" if the room is a space.
If the room does not define a type, the value will be null.
forgotten - Whether all local users have
forgotten the room.
The Room Messages admin API allows server admins to get all messages
sent to a room in a given timeframe. There are various parameters available
that allow for filtering and ordering the returned list. This API supports pagination.
To use it, you will need to authenticate by providing an access_token
for a server admin: see Admin API.
room_id - The ID of the room you wish you fetch messages from.
The following query parameters are available:
from (required) - The token to start returning events from. This token can be obtained from a prev_batch
or next_batch token returned by the /sync endpoint, or from an end token returned by a previous request to this endpoint.
to - The token to stop returning events at.
limit - The maximum number of events to return. Defaults to 10.
filter - A JSON RoomEventFilter to filter returned events with.
dir - The direction to return events from. Either f for forwards or b for backwards. Setting
this value to b will reverse the above sort order. Defaults to f.
Response
The following fields are possible in the JSON response body:
chunk - A list of room events. The order depends on the dir parameter.
Note that an empty chunk does not necessarily imply that no more events are available. Clients should continue to paginate until no end property is returned.
end - A token corresponding to the end of chunk. This token can be passed back to this endpoint to request further events.
If no further events are available, this property is omitted from the response.
start - A token corresponding to the start of chunk.
state - A list of state events relevant to showing the chunk.
The Room Timestamp to Event API endpoint fetches the event_id of the closest event to the given
timestamp (ts query parameter) in the given direction (dir query parameter).
Useful for cases like jump to date so you can start paginating messages from
a given date in the archive.
The API is:
GET /_synapse/admin/v1/rooms/<room_id>/timestamp_to_event
Parameters
The following path parameters are required:
room_id - The ID of the room you wish to check.
The following query parameters are available:
ts - a timestamp in milliseconds where we will find the closest event in
the given direction.
dir - can be f or b to indicate forwards and backwards in time from the
given timestamp. Defaults to f.
Response
event_id - The event ID closest to the given timestamp.
origin_server_ts - The timestamp of the event in milliseconds since the Unix epoch.
The Block Room admin API allows server admins to block and unblock rooms,
and query to see if a given room is blocked.
This API can be used to pre-emptively block a room, even if it's unknown to this
homeserver. Users will be prevented from joining a blocked room.
The following parameters should be set in the URL:
room_id - The ID of the room.
Response
The following fields are possible in the JSON response body:
block - A boolean. true if the room is blocked, otherwise false
user_id - An optional string. If the room is blocked (block is true) shows
the user who has add the room to blocking list. Otherwise it is not displayed.
The Delete Room admin API allows server admins to remove rooms from the server
and block these rooms.
Shuts down a room. Moves all local users and room aliases automatically to a
new room if new_room_user_id is set. Otherwise local users only
leave the room without any information.
The new room will be created with the user specified by the new_room_user_id parameter
as room administrator and will contain a message explaining what happened. Users invited
to the new room will have power level -10 by default, and thus be unable to speak.
If block is true, users will be prevented from joining the old room.
This option can in Version 1 also be used to pre-emptively
block a room, even if it's unknown to this homeserver. In this case, the room will be
blocked, and no further action will be taken. If block is false, attempting to
delete an unknown room is invalid and will be rejected as a bad request.
This API will remove all trace of the old room from your database after removing
all local users. If purge is true (the default), all traces of the old room will
be removed from your database after removing all local users. If you do not want
this to happen, set purge to false.
Depending on the amount of history being purged, a call to the API may take
several minutes or longer.
The local server will only have the power to move local user and room aliases to
the new room. Users on other servers will be unaffected.
This version works synchronously. That means you only get the response once the server has
finished the action, which may take a long time. If you request the same action
a second time, and the server has not finished the first one, the second request will block.
This is fixed in version 2 of this API. The parameters are the same in both APIs.
This API will become deprecated in the future.
The API is:
DELETE /_synapse/admin/v1/rooms/<room_id>
with a body of:
{
"new_room_user_id": "@someuser:example.com",
"room_name": "Content Violation Notification",
"message": "Bad Room has been shutdown due to content violations on this server. Please review our Terms of Service.",
"block": true,
"purge": true
}
Note: This API is new, experimental and "subject to change".
This version works asynchronously, meaning you get the response from server immediately
while the server works on that task in background. You can then request the status of the action
to check if it has completed.
The API is:
DELETE /_synapse/admin/v2/rooms/<room_id>
with a body of:
{
"new_room_user_id": "@someuser:example.com",
"room_name": "Content Violation Notification",
"message": "Bad Room has been shutdown due to content violations on this server. Please review our Terms of Service.",
"block": true,
"purge": true
}
The API starts the shut down and purge running, and returns immediately with a JSON body with
a purge id:
{
"delete_id": "<opaque id>"
}
Parameters
The following parameters should be set in the URL:
room_id - The ID of the room.
The following JSON body parameters are available:
new_room_user_id - Optional. If set, a new room will be created with this user ID
as the creator and admin, and all users in the old room will be moved into that
room. If not set, no new room will be created and the users will just be removed
from the old room. The user ID must be on the local server, but does not necessarily
have to belong to a registered user.
room_name - Optional. A string representing the name of the room that new users will be
invited to. Defaults to Content Violation Notification
message - Optional. A string containing the first message that will be sent as
new_room_user_id in the new room. Ideally this will clearly convey why the
original room was shut down. Defaults to Sharing illegal content on this server is not permitted and rooms in violation will be blocked.
block - Optional. If set to true, this room will be added to a blocking list,
preventing future attempts to join the room. Rooms can be blocked
even if they're not yet known to the homeserver (only with
Version 1 of the API). Defaults to false.
purge - Optional. If set to true, it will remove all traces of the room from your database.
Defaults to true.
force_purge - Optional, and ignored unless purge is true. If set to true, it
will force a purge to go ahead even if there are local users still in the room. Do not
use this unless a regular purge operation fails, as it could leave those users'
clients in a confused state.
The JSON body must not be empty. The body must be at least {}.
Note: This API is new, experimental and "subject to change".
It is possible to query the status of the background task for deleting rooms.
The status can be queried up to 24 hours after completion of the task,
or until Synapse is restarted (whichever happens first).
The following fields are returned in the JSON response body:
results - An array of objects, each containing information about one task.
This field is omitted from the result when you query by delete_id.
Task objects contain the following fields:
delete_id - The ID for this purge if you query by room_id.
status - The status will be one of:
shutting_down - The process is removing users from the room.
purging - The process is purging the room and event data from database.
complete - The process has completed successfully.
failed - The process is aborted, an error has occurred.
error - A string that shows an error message if status is failed.
Otherwise this field is hidden.
shutdown_room - An object containing information about the result of shutting down the room.
Note: The result is shown after removing the room members.
The delete process can still be running. Please pay attention to the status.
kicked_users - An array of users (user_id) that were kicked.
failed_to_kick_users - An array of users (user_id) that that were not kicked.
local_aliases - An array of strings representing the local aliases that were
migrated from the old room to the new.
new_room_id - A string representing the room ID of the new room, or null if
no such room was created.
Note: This guide may be outdated by the time you read it. By nature of room deletions being performed at the database level,
the structure can and does change without notice.
First, it's important to understand that a room deletion is very destructive. Undoing a deletion is not as simple as pretending it
never happened - work has to be done to move forward instead of resetting the past. In fact, in some cases it might not be possible
to recover at all:
If the room was invite-only, your users will need to be re-invited.
If the room no longer has any members at all, it'll be impossible to rejoin.
The first user to rejoin will have to do so via an alias on a different
server (or receive an invite from a user on a different server).
With all that being said, if you still want to try and recover the room:
If the room was blocked, you must unblock it on your server. This can be
accomplished as follows:
For safety reasons, shut down Synapse.
In the database, run DELETE FROM blocked_rooms WHERE room_id = '!example:example.org';
For caution: it's recommended to run this in a transaction: BEGIN; DELETE ...;, verify you got 1 result, then COMMIT;.
The room ID is the same one supplied to the delete room API, not the Content Violation room.
Restart Synapse.
This step is unnecessary if block was not set.
Any room aliases on your server that pointed to the deleted room may have
been deleted, or redirected to the Content Violation room. These will need
to be restored manually.
Users on your server that were in the deleted room will have been kicked
from the room. Consider whether you want to update their membership
(possibly via the Edit Room Membership API) or let
them handle rejoining themselves.
If new_room_user_id was given, a 'Content Violation' will have been
created. Consider whether you want to delete that roomm.
Grants another user the highest power available to a local user who is in the room.
If the user is not in the room, and it is not publicly joinable, then invite the user.
By default the server admin (the caller) is granted power, but another user can
optionally be specified, e.g.:
POST /_synapse/admin/v1/rooms/<room_id_or_alias>/make_room_admin
{
"user_id": "@foo:example.com"
}
Enables querying and deleting forward extremities from rooms. When a lot of forward
extremities accumulate in a room, performance can become degraded. For details, see
#1760.
WARNING: Please ensure you know what you're doing and have read
the related issue #1760.
Under no situations should this API be executed as an automated maintenance task!
If a room has lots of forward extremities, the extra can be
deleted as follows:
You will need to authenticate with an access token for an admin user.
When using the PUT form, retransmissions with the same transaction ID will be
ignored in the same way as with PUT /_matrix/client/r0/rooms/{roomId}/send/{eventType}/{txnId}.
The request body should look something like the following:
{
"user_id": "@target_user:server_name",
"content": {
"msgtype": "m.text",
"body": "This is my message"
}
}
You can optionally include the following additional parameters:
type: the type of event. Defaults to m.room.message.
state_key: Setting this will result in a state event being sent.
Once the notice has been sent, the API will return the following response:
{
"event_id": "<event_id>"
}
Note that server notices must be enabled in homeserver.yaml before this API
can be used. See the server notices documentation for more information.
To paginate, check for next_token and if present, call the endpoint
again with from set to the value of next_token. This will return a new page.
If the endpoint does not return a next_token then there are no more
reports to paginate through.
Parameters
The following parameters should be set in the URL:
limit: string representing a positive integer - Is optional but is
used for pagination, denoting the maximum number of items to return
in this call. Defaults to 100.
from: string representing a positive integer - Is optional but used for pagination,
denoting the offset in the returned results. This should be treated as an opaque value
and not explicitly set to anything other than the return value of next_token from a
previous call. Defaults to 0.
order_by - string - The method in which to sort the returned list of users. Valid values are:
user_id - Users are ordered alphabetically by user_id. This is the default.
displayname - Users are ordered alphabetically by displayname.
media_length - Users are ordered by the total size of uploaded media in bytes.
Smallest to largest.
media_count - Users are ordered by number of uploaded media. Smallest to largest.
from_ts - string representing a positive integer - Considers only
files created at this timestamp or later. Unix timestamp in ms.
until_ts - string representing a positive integer - Considers only
files created at this timestamp or earlier. Unix timestamp in ms.
search_term - string - Filter users by their user ID localpart or displayname.
The search term can be found in any part of the string.
Defaults to no filtering.
dir - string - Direction of order. Either f for forwards or b for backwards.
Setting this value to b will reverse the above sort order. Defaults to f.
Response
The following fields are returned in the JSON response body:
users - An array of objects, each containing information
about the user and their local media. Objects contain the following fields:
displayname - string - Displayname of this user.
media_count - integer - Number of uploaded media by this user.
media_length - integer - Size of uploaded media in bytes by this user.
user_id - string - Fully-qualified user ID (ex. @user:server.com).
next_token - integer - Opaque value used for pagination. See above.
total - integer - Total number of users after filtering.
Returns the 10 largest rooms and an estimate of how much space in the database
they are taking.
This does not include the size of any associated media associated with the room.
Returns an error on SQLite.
Note: This uses the planner statistics from PostgreSQL to do the estimates,
which means that the returned information can vary widely from reality. However,
it should be enough to get a rough idea of where database disk space is going.
user_id - A fully-qualified user id. For example, @user:server.com.
Body parameters:
password - string, optional. If provided, the user's password is updated and all
devices are logged out, unless logout_devices is set to false.
logout_devices - bool, optional, defaults to true. If set to false, devices aren't
logged out even when password is provided.
displayname - string, optional. If set to an empty string (""), the user's display name
will be removed.
avatar_url - string, optional. Must be a
MXC URI.
If set to an empty string (""), the user's avatar is removed.
threepids - array, optional. If provided, the user's third-party IDs (email, msisdn) are
entirely replaced with the given list. Each item in the array is an object with the following
fields:
medium - string, required. The type of third-party ID, either email or msisdn (phone number).
address - string, required. The third-party ID itself, e.g. alice@example.com for email or
447470274584 (for a phone number with country code "44") and 19254857364 (for a phone number
with country code "1") for msisdn.
Note: If a threepid is removed from a user via this option, Synapse will also attempt to remove
that threepid from any identity servers it is aware has a binding for it.
external_ids - array, optional. Allow setting the identifier of the external identity
provider for SSO (Single sign-on). More details are in the configuration manual under the
sections sso and oidc_providers.
auth_provider - string, required. The unique, internal ID of the external identity provider.
The same as idp_id from the homeserver configuration. Note that no error is raised if the
provided value is not in the homeserver configuration.
external_id - string, required. An identifier for the user in the external identity provider.
When the user logs in to the identity provider, this must be the unique ID that they map to.
admin - bool, optional, defaults to false. Whether the user is a homeserver administrator,
granting them access to the Admin API, among other things.
deactivated - bool, optional. If unspecified, deactivation state will be left unchanged.
Note: the password field must also be set if both of the following are true:
deactivated is set to false and the user was previously deactivated (you are reactivating this user)
Users are allowed to set their password on this homeserver (both password_config.enabled and
password_config.localdb_enabled config options are set to true).
Users' passwords are wiped upon account deactivation, hence the need to set a new one here.
Note: a user cannot be erased with this API. For more details on
deactivating and erasing users see Deactivate Account.
locked - bool, optional. If unspecified, locked state will be left unchanged.
user_type - string or null, optional. If not provided, the user type will be
not be changed. If null is given, the user type will be cleared.
Other allowed options are: bot and support.
To paginate, check for next_token and if present, call the endpoint again
with from set to the value of next_token. This will return a new page.
If the endpoint does not return a next_token then there are no more users
to paginate through.
Parameters
The following parameters should be set in the URL:
user_id - Is optional and filters to only return users with user IDs
that contain this value. This parameter is ignored when using the name parameter.
name - Is optional and filters to only return users with user ID localparts
or displaynames that contain this value.
guests - string representing a bool - Is optional and if false will exclude guest users.
Defaults to true to include guest users. This parameter is not supported when MSC3861 is enabled. See #15582
admins - Optional flag to filter admins. If true, only admins are queried. If false, admins are excluded from
the query. When the flag is absent (the default), both admins and non-admins are included in the search results.
deactivated - string representing a bool - Is optional and if true will include deactivated users.
Defaults to false to exclude deactivated users.
limit - string representing a positive integer - Is optional but is used for pagination,
denoting the maximum number of items to return in this call. Defaults to 100.
from - string representing a positive integer - Is optional but used for pagination,
denoting the offset in the returned results. This should be treated as an opaque value and
not explicitly set to anything other than the return value of next_token from a previous call.
Defaults to 0.
order_by - The method by which to sort the returned list of users.
If the ordered field has duplicates, the second order is always by ascending name,
which guarantees a stable ordering. Valid values are:
name - Users are ordered alphabetically by name. This is the default.
is_guest - Users are ordered by is_guest status.
admin - Users are ordered by admin status.
user_type - Users are ordered alphabetically by user_type.
deactivated - Users are ordered by deactivated status.
shadow_banned - Users are ordered by shadow_banned status.
displayname - Users are ordered alphabetically by displayname.
avatar_url - Users are ordered alphabetically by avatar URL.
creation_ts - Users are ordered by when the users was created in ms.
last_seen_ts - Users are ordered by when the user was lastly seen in ms.
dir - Direction of media order. Either f for forwards or b for backwards.
Setting this value to b will reverse the above sort order. Defaults to f.
not_user_type - Exclude certain user types, such as bot users, from the request.
Can be provided multiple times. Possible values are bot, support or "empty string".
"empty string" here means to exclude users without a type.
locked - string representing a bool - Is optional and if true will include locked users.
Defaults to false to exclude locked users. Note: Introduced in v1.93.
Caution. The database only has indexes on the columns name and creation_ts.
This means that if a different sort order is used (is_guest, admin,
user_type, deactivated, shadow_banned, avatar_url or displayname),
this can cause a large load on the database, especially for large environments.
Response
The following fields are returned in the JSON response body:
users - An array of objects, each containing information about an user.
User objects contain the following fields:
name - string - Fully-qualified user ID (ex. @user:server.com).
is_guest - bool - Status if that user is a guest account.
admin - bool - Status if that user is a server administrator.
user_type - string - Type of the user. Normal users are type None.
This allows user type specific behaviour. There are also types support and bot.
deactivated - bool - Status if that user has been marked as deactivated.
erased - bool - Status if that user has been marked as erased.
shadow_banned - bool - Status if that user has been marked as shadow banned.
displayname - string - The user's display name if they have set one.
avatar_url - string - The user's avatar URL if they have set one.
creation_ts - integer - The user's creation timestamp in ms.
last_seen_ts - integer - The user's last activity timestamp in ms.
locked - bool - Status if that user has been marked as locked. Note: Introduced in v1.93.
next_token: string representing a positive integer - Indication for pagination. See above.
total - integer - Total number of media.
Added in Synapse 1.93: the locked query parameter and response field.
This API deactivates an account. It removes active access tokens, resets the
password, and deletes third-party IDs (to prevent the user requesting a
password reset).
It can also mark the user as GDPR-erased. This means messages sent by the
user will still be visible by anyone that was in the room when these messages
were sent, but hidden from users joining the room afterwards.
The api is:
POST /_synapse/admin/v1/deactivate/<user_id>
with a body of:
{
"erase": true
}
The erase parameter is optional and defaults to false.
An empty body may be passed for backwards compatibility.
The following actions are performed when deactivating an user:
Try to unbind 3PIDs from the identity server
Remove all 3PIDs from the homeserver
Delete all devices and E2EE keys
Delete all access tokens
Delete all pushers
Delete the password hash
Removal from all rooms the user is a member of
Remove the user from the user directory
Reject all pending invites
Remove all account validity information related to the user
Remove the arbitrary data store known as account data. For example, this includes:
list of ignored users;
push rules;
secret storage keys; and
cross-signing keys.
The following additional actions are performed during deactivation if erase
is set to true:
Remove the user's display name
Remove the user's avatar URL
Mark the user as erased
The following actions are NOT performed. The list may be incomplete.
The server returns the list of rooms of which the user and the server
are member. If the user is local, all the rooms of which the user is
member are returned.
Parameters
The following parameters should be set in the URL:
user_id - fully qualified: for example, @user:server.com.
Response
The following fields are returned in the JSON response body:
Gets a list of all local media that a specific user_id has created.
These are media that the user has uploaded themselves
(local media), as well as
URL preview images requested by the user if the
feature is enabled.
By default, the response is ordered by descending creation date and ascending media ID.
The newest media is on top. You can change the order with parameters
order_by and dir.
To paginate, check for next_token and if present, call the endpoint again
with from set to the value of next_token. This will return a new page.
If the endpoint does not return a next_token then there are no more
reports to paginate through.
Parameters
The following parameters should be set in the URL:
user_id - string - fully qualified: for example, @user:server.com.
limit: string representing a positive integer - Is optional but is used for pagination,
denoting the maximum number of items to return in this call. Defaults to 100.
from: string representing a positive integer - Is optional but used for pagination,
denoting the offset in the returned results. This should be treated as an opaque value and
not explicitly set to anything other than the return value of next_token from a previous call.
Defaults to 0.
order_by - The method by which to sort the returned list of media.
If the ordered field has duplicates, the second order is always by ascending media_id,
which guarantees a stable ordering. Valid values are:
media_id - Media are ordered alphabetically by media_id.
upload_name - Media are ordered alphabetically by name the media was uploaded with.
created_ts - Media are ordered by when the content was uploaded in ms.
Smallest to largest. This is the default.
last_access_ts - Media are ordered by when the content was last accessed in ms.
Smallest to largest.
media_length - Media are ordered by length of the media in bytes.
Smallest to largest.
media_type - Media are ordered alphabetically by MIME-type.
quarantined_by - Media are ordered alphabetically by the user ID that
initiated the quarantine request for this media.
safe_from_quarantine - Media are ordered by the status if this media is safe
from quarantining.
dir - Direction of media order. Either f for forwards or b for backwards.
Setting this value to b will reverse the above sort order. Defaults to f.
If neither order_by nor dir is set, the default order is newest media on top
(corresponds to order_by = created_ts and dir = b).
Caution. The database only has indexes on the columns media_id,
user_id and created_ts. This means that if a different sort order is used
(upload_name, last_access_ts, media_length, media_type,
quarantined_by or safe_from_quarantine), this can cause a large load on the
database, especially for large environments.
Response
The following fields are returned in the JSON response body:
media - An array of objects, each containing information about a media.
Media objects contain the following fields:
created_ts - integer - Timestamp when the content was uploaded in ms.
last_access_ts - integer or null - Timestamp when the content was last accessed in ms.
Null if there was no access, yet.
media_id - string - The id used to refer to the media. Details about the format
are documented under
media repository.
media_length - integer - Length of the media in bytes.
media_type - string - The MIME-type of the media.
quarantined_by - string or null - The user ID that initiated the quarantine request
for this media. Null if not quarantined.
safe_from_quarantine - bool - Status if this media is safe from quarantining.
upload_name - string or null - The name the media was uploaded with. Null if not provided during upload.
next_token: integer - Indication for pagination. See above.
This API deletes the local media from the disk of your own server
that a specific user_id has created. This includes any local thumbnails.
This API will not affect media that has been uploaded to external
media repositories (e.g https://github.com/turt2live/matrix-media-repo/).
By default, the API deletes media ordered by descending creation date and ascending media ID.
The newest media is deleted first. You can change the order with parameters
order_by and dir. If no limit is set the API deletes 100 files per request.
The following fields are returned in the JSON response body:
deleted_media: an array of strings - List of deleted media_id
total: integer - Total number of deleted media_id
Note: There is no next_token. This is not useful for deleting media, because
after deleting media the remaining media have a new order.
Parameters
This API has the same parameters as
List media uploaded by a user.
With the parameters you can for example limit the number of files to delete at once or
delete largest/smallest or newest/oldest files first.
Note: This API is disabled when MSC3861 is enabled. See #15582
Get an access token that can be used to authenticate as that user. Useful for
when admins wish to do actions on behalf of a user.
The API is:
POST /_synapse/admin/v1/users/<user_id>/login
{}
An optional valid_until_ms field can be specified in the request body as an
integer timestamp that specifies when the token should expire. By default tokens
do not expire. Note that this API does not allow a user to login as themselves
(to create more tokens).
This API does not generate a new device for the user, and so will not appear
their /devices list, and in general the target user should not be able to
tell they have been logged in as.
To expire the token call the standard /logout API with the token.
Note: The token will expire if the admin user calls /logout/all from any
of their devices, but the token will not expire if the target user does the
same.
This endpoint is not intended for server administrator usage;
we describe it here for completeness.
This API temporarily permits a user to replace their master cross-signing key
without going through
user-interactive authentication (UIA).
This is useful when Synapse has delegated its authentication to the
Matrix Authentication Service;
as Synapse cannot perform UIA is not possible in these circumstances.
The API is
POST /_synapse/admin/v1/users/<user_id>/_allow_cross_signing_replacement_without_uia
{}
If the user does not exist, or does exist but has no master cross-signing key,
this will return with status code 404 Not Found.
Otherwise, a response body like the following is returned, with status 200 OK:
{
"updatable_without_uia_before_ms": 1234567890
}
The response body is a JSON object with a single field:
updatable_without_uia_before_ms: integer. The timestamp in milliseconds
before which the user is permitted to replace their cross-signing key without
going through UIA.
The following parameters should be set in the URL:
user_id - fully qualified: for example, @user:server.com.
Response
The following fields are returned in the JSON response body:
devices - An array of objects, each containing information about a device.
Device objects contain the following fields:
device_id - Identifier of device.
display_name - Display name set by the user for this device.
Absent if no name has been set.
last_seen_ip - The IP address where this device was last seen.
(May be a few minutes out of date, for efficiency reasons).
last_seen_user_agent - The user agent of the device when it was last seen.
(May be a few minutes out of date, for efficiency reasons).
last_seen_ts - The timestamp (in milliseconds since the unix epoch) when this
devices was last seen. (May be a few minutes out of date, for efficiency reasons).
The following parameters should be set in the URL:
user_id - fully qualified: for example, @user:server.com.
device_id - The device to retrieve.
Response
The following fields are returned in the JSON response body:
device_id - Identifier of device.
display_name - Display name set by the user for this device.
Absent if no name has been set.
last_seen_ip - The IP address where this device was last seen.
(May be a few minutes out of date, for efficiency reasons).
last_seen_user_agent - The user agent of the device when it was last seen.
(May be a few minutes out of date, for efficiency reasons).
last_seen_ts - The timestamp (in milliseconds since the unix epoch) when this
devices was last seen. (May be a few minutes out of date, for efficiency reasons).
Shadow-banning is a useful tool for moderating malicious or egregiously abusive users.
A shadow-banned users receives successful responses to their client-server API requests,
but the events are not propagated into rooms. This can be an effective tool as it
(hopefully) takes longer for the user to realise they are being moderated before
pivoting to another account.
Shadow-banning a user should be used as a tool of last resort and may lead to confusing
or broken behaviour for the client. A shadow-banned user will not receive any
notification and it is generally more appropriate to ban or kick abusive users.
A shadow-banned user will be unable to contact anyone on the server.
To shadow-ban a user the API is:
POST /_synapse/admin/v1/users/<user_id>/shadow_ban
GET /_synapse/admin/v1/auth_providers/$provider/users/$external_id
When a user matched the given ID for the given provider, an HTTP code 200 with a response body like the following is returned:
{
"user_id": "@hello:example.org"
}
Parameters
The following parameters should be set in the URL:
provider - The ID of the authentication provider, as advertised by the GET /_matrix/client/v3/login API in the m.login.sso authentication method.
external_id - The user ID from the authentication provider. Usually corresponds to the sub claim for OIDC providers, or to the uid attestation for SAML2 providers.
The external_id may have characters that are not URL-safe (typically /, : or @), so it is advised to URL-encode those parameters.
Errors
Returns a 404 HTTP status code if no user was found, with a response body like this:
{
"errcode":"M_NOT_FOUND",
"error":"User not found"
}
This API returns the running Synapse version.
This is useful when a Synapse instance
is behind a proxy that does not forward the 'Server' header (which also
contains Synapse version information).
This API gets the current destination retry timing info for all remote servers.
The list contains all the servers with which the server federates,
regardless of whether an error occurred or not.
If an error occurs, it may take up to 20 minutes for the error to be displayed here,
as a complete retry must have failed.
To paginate, check for next_token and if present, call the endpoint again
with from set to the value of next_token. This will return a new page.
If the endpoint does not return a next_token then there are no more destinations
to paginate through.
Parameters
The following query parameters are available:
from - Offset in the returned list. Defaults to 0.
limit - Maximum amount of destinations to return. Defaults to 100.
order_by - The method in which to sort the returned list of destinations.
Valid values are:
destination - Destinations are ordered alphabetically by remote server name.
This is the default.
retry_last_ts - Destinations are ordered by time of last retry attempt in ms.
retry_interval - Destinations are ordered by how long until next retry in ms.
failure_ts - Destinations are ordered by when the server started failing in ms.
last_successful_stream_ordering - Destinations are ordered by the stream ordering
of the most recent successfully-sent PDU.
dir - Direction of room order. Either f for forwards or b for backwards. Setting
this value to b will reverse the above sort order. Defaults to f.
Caution: The database only has an index on the column destination.
This means that if a different sort order is used,
this can cause a large load on the database, especially for large environments.
Response
The following fields are returned in the JSON response body:
destinations - An array of objects, each containing information about a destination.
Destination objects contain the following fields:
destination - string - Name of the remote server to federate.
retry_last_ts - integer - The last time Synapse tried and failed to reach the
remote server, in ms. This is 0 if the last attempt to communicate with the
remote server was successful.
retry_interval - integer - How long since the last time Synapse tried to reach
the remote server before trying again, in ms. This is 0 if no further retrying occurring.
failure_ts - nullable integer - The first time Synapse tried and failed to reach the
remote server, in ms. This is null if communication with the remote server has never failed.
last_successful_stream_ordering - nullable integer - The stream ordering of the most
recent successfully-sent PDU
to this destination, or null if this information has not been tracked yet.
next_token: string representing a positive integer - Indication for pagination. See above.
Synapse makes federation requests to other homeservers. If a federation request fails,
Synapse will mark the destination homeserver as offline, preventing any future requests
to that server for a "cooldown" period. This period grows over time if the server
continues to fail its responses
(exponential backoff).
Admins can cancel the cooldown period with this API.
This API resets the retry timing for a specific remote server and tries to connect to
the remote server again. It does not wait for the next retry_interval.
The connection must have previously run into an error and retry_last_ts
(Destination Details API) must not be equal to 0.
The connection attempt is carried out in the background and can take a while
even if the API already returns the http status 200.
The API is:
POST /_synapse/admin/v1/federation/destinations/<destination>/reset_connection
{}
Parameters
The following parameters should be set in the URL:
The "manhole" allows server administrators to access a Python shell on a running
Synapse installation. This is a very powerful mechanism for administration and
debugging.
Security Warning
Note that this will give administrative access to synapse to all users with
shell access to the server. It should therefore not be enabled in
environments where untrusted users have shell access.
To enable it, first add the manhole listener configuration in your
homeserver.yaml. You can find information on how to do that
in the configuration manual.
The configuration is slightly different if you're using docker.
When using docker run to start the server, you will then need to change the command to the following to include the
manhole port forwarding. The -p 127.0.0.1:9000:9000 below is important: it
ensures that access to the manhole is only possible for local users.
If you are not using docker, set bind_addresses to ['::1', '127.0.0.1'] as shown.
The bind_addresses in the example below is important: it ensures that access to the
manhole is only possible for local users).
Then restart synapse, and point an ssh client at port 9000 on localhost, using
the username and password configured in homeserver.yaml - with the default
configuration, this would be:
ssh -p9000 matrix@localhost
Then enter the password when prompted (the default is rabbithole).
This gives a Python REPL in which hs gives access to the
synapse.server.HomeServer object - which in turn gives access to many other
parts of the process.
Note that, prior to Synapse 1.41, any call which returns a coroutine will need to be wrapped in ensureDeferred.
As a simple example, retrieving an event from the database:
>>> from twisted.internet import defer
>>> defer.ensureDeferred(hs.get_datastores().main.get_event('$1416420717069yeQaw:matrix.org'))
<Deferred at 0x7ff253fc6998 current result: <FrozenEvent event_id='$1416420717069yeQaw:matrix.org', type='m.room.create', state_key=''>>
In homeserver.yaml, make sure enable_metrics is
set to True.
Enable the /_synapse/metrics Synapse endpoint that Prometheus uses to
collect data:
There are two methods of enabling the metrics endpoint in Synapse.
The first serves the metrics as a part of the usual web server and
can be enabled by adding the metrics resource to the existing
listener as such as in this example:
listeners:
- port: 8008
tls: false
type: http
x_forwarded: true
bind_addresses: ['::1', '127.0.0.1']
resources:
# added "metrics" in this line
- names: [client, federation, metrics]
compress: false
This provides a simple way of adding metrics to your Synapse
installation, and serves under /_synapse/metrics. If you do not
wish your metrics be publicly exposed, you will need to either
filter it out at your load balancer, or use the second method.
The second method runs the metrics server on a different port, in a
different thread to Synapse. This can make it more resilient to
heavy load meaning metrics cannot be retrieved, and can be exposed
to just internal networks easier. The served metrics are available
over HTTP only, and will be available at /_synapse/metrics.
Add a new listener to homeserver.yaml as in this example:
To monitor a Synapse installation using workers,
every worker needs to be monitored independently, in addition to
the main homeserver process. This is because workers don't send
their metrics to the main homeserver process, but expose them
directly (if they are configured to do so).
To allow collecting metrics from a worker, you need to add a
metrics listener to its configuration, by adding the following
under worker_listeners:
- type: metrics
bind_address: ''
port: 9101
The bind_address and port parameters should be set so that
the resulting listener can be reached by prometheus, and they
don't clash with an existing worker.
With this example, the worker's metrics would then be available
on http://127.0.0.1:9101.
Example Prometheus target for Synapse with workers:
Synapse 1.2 updates the Prometheus metrics to match the naming
convention of the upstream prometheus_client. The old names are
considered deprecated and will be removed in a future version of
Synapse.
The old names will be disabled by default in Synapse v1.71.0 and removed
altogether in Synapse v1.73.0.
The duplicated metrics deprecated in Synapse 0.27.0 have been removed.
All time duration-based metrics have been changed to be seconds. This
affects:
msec -> sec metrics
python_gc_time
python_twisted_reactor_tick_time
synapse_storage_query_time
synapse_storage_schedule_time
synapse_storage_transaction_time
Several metrics have been changed to be histograms, which sort entries
into buckets and allow better analysis. The following metrics are now
histograms:
Synapse 0.27.0 begins the process of rationalising the duplicate
*:count metrics reported for the resource tracking for code blocks and
HTTP requests.
At the same time, the corresponding *:total metrics are being renamed,
as the :total suffix no longer makes sense in the absence of a
corresponding :count metric.
To enable a graceful migration path, this release just adds new names
for the metrics being renamed. A future release will remove the old
ones.
The following table shows the new metrics, and the old metrics which
they are replacing.
As of synapse version 0.18.2, the format of the process-wide metrics has
been changed to fit prometheus standard naming conventions. Additionally
the units have been changed to seconds, from milliseconds.
New name
Old name
process_cpu_user_seconds_total
process_resource_utime / 1000
process_cpu_system_seconds_total
process_resource_stime / 1000
process_open_fds (no 'type' label)
process_fds
The python-specific counts of garbage collector performance have been
renamed.
New name
Old name
python_gc_time
reactor_gc_time
python_gc_unreachable_total
reactor_gc_unreachable
python_gc_counts
reactor_gc_counts
The twisted-specific reactor metrics have been renamed.
When generating your Synapse configuration file, you are asked whether you
would like to report usage statistics to Matrix.org. These statistics
provide the foundation a glimpse into the number of Synapse homeservers
participating in the network, as well as statistics such as the number of
rooms being created and messages being sent. This feature is sometimes
affectionately called "phone home" stats. Reporting
is optional
and the reporting endpoint
can be configured,
in case you would like to instead report statistics from a set of homeservers
to your own infrastructure.
This documentation aims to define the statistics available and the
homeserver configuration options that exist to tweak it.
The following statistics are sent to the configured reporting endpoint:
Statistic Name
Type
Description
homeserver
string
The homeserver's server name.
memory_rss
int
The memory usage of the process (in kilobytes on Unix-based systems, bytes on MacOS).
cpu_average
int
CPU time in % of a single core (not % of all cores).
server_context
string
An arbitrary string used to group statistics from a set of homeservers.
timestamp
int
The current time, represented as the number of seconds since the epoch.
uptime_seconds
int
The number of seconds since the homeserver was last started.
python_version
string
The Python version number in use (e.g "3.7.1"). Taken from sys.version_info.
total_users
int
The number of registered users on the homeserver.
total_nonbridged_users
int
The number of users, excluding those created by an Application Service.
daily_user_type_native
int
The number of native users created in the last 24 hours.
daily_user_type_guest
int
The number of guest users created in the last 24 hours.
daily_user_type_bridged
int
The number of users created by Application Services in the last 24 hours.
total_room_count
int
The total number of rooms present on the homeserver.
daily_active_users
int
The number of unique users1 that have used the homeserver in the last 24 hours.
monthly_active_users
int
The number of unique users1 that have used the homeserver in the last 30 days.
daily_active_rooms
int
The number of rooms that have had a (state) event with the type m.room.message sent in them in the last 24 hours.
daily_active_e2ee_rooms
int
The number of rooms that have had a (state) event with the type m.room.encrypted sent in them in the last 24 hours.
daily_messages
int
The number of (state) events with the type m.room.message seen in the last 24 hours.
daily_e2ee_messages
int
The number of (state) events with the type m.room.encrypted seen in the last 24 hours.
daily_sent_messages
int
The number of (state) events sent by a local user with the type m.room.message seen in the last 24 hours.
daily_sent_e2ee_messages
int
The number of (state) events sent by a local user with the type m.room.encrypted seen in the last 24 hours.
r30v2_users_all
int
The number of 30 day retained users, with a revised algorithm. Defined as users that appear more than once in the past 60 days, and have more than 30 days between the most and least recent appearances in the past 60 days. Includes clients that do not fit into the below r30 client types.
r30v2_users_android
int
The number of 30 day retained users, as defined above. Filtered only to clients with ("riot" or "element") and "android" (case-insensitive) in the user agent string.
r30v2_users_ios
int
The number of 30 day retained users, as defined above. Filtered only to clients with ("riot" or "element") and "ios" (case-insensitive) in the user agent string.
r30v2_users_electron
int
The number of 30 day retained users, as defined above. Filtered only to clients with ("riot" or "element") and "electron" (case-insensitive) in the user agent string.
r30v2_users_web
int
The number of 30 day retained users, as defined above. Filtered only to clients with "mozilla" or "gecko" (case-insensitive) in the user agent string.
The database engine that is in use. Either "psycopg2" meaning PostgreSQL is in use, or "sqlite3" for SQLite3.
database_server_version
string
The version of the database server. Examples being "10.10" for PostgreSQL server version 10.0, and "3.38.5" for SQLite 3.38.5 installed on the system.
log_level
string
The log level in use. Examples are "INFO", "WARNING", "ERROR", "DEBUG", etc.
1
Native matrix users and guests are always counted. If the
track_puppeted_user_ips
option is set to true, "puppeted" users (users that an Application Service have performed
an action on behalf of)
will also be counted. Note that an Application Service can "puppet" any user in their
user namespace,
not only users that the Application Service has created. If this happens, the Application Service
will additionally be counted as a user (irrespective of track_puppeted_user_ips).
If statistics reporting is enabled, the endpoint that Synapse sends metrics to is configured by the
report_stats_endpoint config
option. By default, statistics are sent to Matrix.org.
If you would like to set up your own statistics collection server and send metrics there, you may
consider using one of the following known implementations:
Synapse can be configured to record the number of monthly active users (also referred to as MAU) on a given homeserver.
For clarity's sake, MAU only tracks local users.
Please note that the metrics recorded by the Homeserver Usage Stats
are calculated differently. The monthly_active_users from the usage stats does not take into account any
of the rules below, and counts any users who have made a request to the homeserver in the last 30 days.
Individual user activity is measured in active days. If a user performs an action, the exact time of that action is then recorded. When
calculating the MAU figure, any users with a recorded action in the last 30 days are considered part of the cohort. Days are measured
as a rolling window from the current system time to 30 days ago.
So for example, if Synapse were to calculate the active users on the 15th July at 13:25, it would include any activity from 15th June 13:25 onwards.
A user is never considered active if they are either:
Part of the trial day cohort (described below)
Owned by an application service.
Note: This only covers users that are part of an application service namespaces.users registration. The namespace
must also be marked as exclusive.
Otherwise, any request to Synapse will mark the user as active. Please note that registration will not mark a user as active unless
they register with a 3pid that is included in the config field mau_limits_reserved_threepids.
The Prometheus metric for MAU is refreshed every 5 minutes.
Once an hour, Synapse checks to see if any users are inactive (with only activity timestamps later than 30 days). These users
are removed from the active users cohort. If they then become active, they are immediately restored to the cohort.
It is important to note that deactivated users are not immediately removed from the pool of active users, but as these users won't
perform actions they will eventually be removed from the cohort.
If the config option mau_trial_days is set, a user must have been active this many days after registration to be active. A user is in the
trial period if their registration timestamp (also known as the creation_ts) is less than mau_trial_days old.
As an example, if mau_trial_days is set to 3 and a user is active after 3 days (72 hours from registration time) then they will be counted as active.
The mau_appservice_trial_days config further extends this rule by applying different durations depending on the appservice_id of the user.
Users registered by an application service will be recorded with an appservice_id matching the id key in the registration file for that service.
If both config options limit_usage_by_mau and max_mau_value is set, and the current MAU value exceeds the maximum value, the
homeserver will begin to block some actions.
Individual users matching any of the below criteria never have their actions blocked:
Considered part of the cohort of MAU users.
Considered part of the trial period.
Registered as a support user.
Application service users if track_appservice_user_ips is NOT set.
Please not that server admins are not exempt from blocking.
The following actions are blocked when the MAU limit is exceeded:
Logging in
Sending events
Creating rooms
Syncing
Registration is also blocked for all new signups unless the user is registering with a threepid included in the mau_limits_reserved_threepids
config value.
When a request is blocked, the response will have the errcodeM_RESOURCE_LIMIT_EXCEEDED.
Synapse records several different prometheus metrics for MAU.
synapse_admin_mau_current records the current MAU figure for native (non-application-service) users.
synapse_admin_mau_max records the maximum MAU as dictated by the max_mau_value config value.
synapse_admin_mau_current_mau_by_service records the current MAU including application service users. The label app_service can be used
to filter by a specific service ID. This also includes non-application-service users under app_service=native .
synapse_admin_mau_registered_reserved_users records the number of users specified in mau_limits_reserved_threepids which have
registered accounts on the homeserver.
It is possible to monitor much of the internal state of Synapse using Prometheus
metrics and Grafana.
A guide for configuring Synapse to provide metrics is available here
and information on setting up Grafana is here.
In this setup, Prometheus will periodically scrape the information Synapse provides and
store a record of it over time. Grafana is then used as an interface to query and
present this information through a series of pretty graphs.
Once you have grafana set up, and assuming you're using our grafana dashboard template, look for the following graphs when debugging a slow/overloaded Synapse:
This, along with the CPU and Memory graphs, is a good way to check the general health of your Synapse instance. It represents how long it takes for a user on your homeserver to send a message.
These graphs show the database transactions that are occurring the most frequently, as well as those are that are taking the most amount of time to execute.
In the first graph, we can see obvious spikes corresponding to lots of get_user_by_id transactions. This would be useful information to figure out which part of the Synapse codebase is potentially creating a heavy load on the system. However, be sure to cross-reference this with Transaction Duration, which states that get_users_by_id is actually a very quick database transaction and isn't causing as much load as others, like persist_events:
Still, it's probably worth investigating why we're getting users from the database that often, and whether it's possible to reduce the amount of queries we make by adjusting our cache factor(s).
The persist_events transaction is responsible for saving new room events to the Synapse database, so can often show a high transaction duration.
The charts in the "Federation" section show information about incoming and outgoing federation requests. Federation data can be divided into two basic types:
PDU (Persistent Data Unit) - room events: messages, state events (join/leave), etc. These are permanently stored in the database.
EDU (Ephemeral Data Unit) - other data, which need not be stored permanently, such as read receipts, typing notifications.
The "Outgoing EDUs by type" chart shows the EDUs within outgoing federation requests by type: m.device_list_update, m.direct_to_device, m.presence, m.receipt, m.typing.
If you see a large number of m.presence EDUs and are having trouble with too much CPU load, you can disable presence in the Synapse config. See also #3971.
This is quite a useful graph. It shows how many times Synapse attempts to retrieve a piece of data from a cache which the cache did not contain, thus resulting in a call to the database. We can see here that the _get_joined_profile_from_event_id cache is being requested a lot, and often the data we're after is not cached.
Cross-referencing this with the Eviction Rate graph, which shows that entries are being evicted from _get_joined_profile_from_event_id quite often:
we should probably consider raising the size of that cache by raising its cache factor (a multiplier value for the size of an individual cache). Information on doing so is available here (note that the configuration of individual cache factors through the configuration file is available in Synapse v1.14.0+, whereas doing so through environment variables has been supported for a very long time). Note that this will increase Synapse's overall memory usage.
Forward extremities are the leaf events at the end of a DAG in a room, aka events that have no children. The more that exist in a room, the more state resolution that Synapse needs to perform (hint: it's an expensive operation). While Synapse has code to prevent too many of these existing at one time in a room, bugs can sometimes make them crop up again.
If a room has >10 forward extremities, it's worth checking which room is the culprit and potentially removing them using the SQL queries mentioned in #1760.
Large spikes in garbage collection times (bigger than shown here, I'm talking in the
multiple seconds range), can cause lots of problems in Synapse performance. It's more an
indicator of problems, and a symptom of other problems though, so check other graphs for what might be causing it.
If you're still having performance problems with your Synapse instance and you've
tried everything you can, it may just be a lack of system resources. Consider adding
more CPU and RAM, and make use of worker mode
to make use of multiple CPU cores / multiple machines for your homeserver.
SELECT nspname || '.' || relname AS "relation",
pg_size_pretty(pg_total_relation_size(c.oid)) AS "total_size"
FROM pg_class c
LEFT JOIN pg_namespace n ON (n.oid = c.relnamespace)
WHERE nspname NOT IN ('pg_catalog', 'information_schema')
AND c.relkind <> 'i'
AND nspname !~ '^pg_toast'
ORDER BY pg_total_relation_size(c.oid) DESC
LIMIT 20;
You get the same information when you use the
admin API
and set parameter order_by=state_events.
SELECT r.name, s.room_id, s.current_state_events
FROM room_stats_current s
LEFT JOIN room_stats_state r USING (room_id)
ORDER BY current_state_events DESC
LIMIT 20;
and by state_group_events count:
SELECT rss.name, s.room_id, COUNT(s.room_id)
FROM state_groups_state s
LEFT JOIN room_stats_state rss USING (room_id)
GROUP BY s.room_id, rss.name
ORDER BY COUNT(s.room_id) DESC
LIMIT 20;
plus same, but with join removed for performance reasons:
SELECT s.room_id, COUNT(s.room_id)
FROM state_groups_state s
GROUP BY s.room_id
ORDER BY COUNT(s.room_id) DESC
LIMIT 20;
SELECT e.room_id, r.name, COUNT(e.event_id) cnt
FROM events e
LEFT JOIN room_stats_state r USING (room_id)
WHERE e.origin_server_ts >= DATE_PART('epoch', NOW() - INTERVAL '1 day') * 1000
GROUP BY e.room_id, r.name
ORDER BY cnt DESC
LIMIT 20;
Caution. This query does not use any indexes, can be slow and create load on the database.
SELECT COUNT(*), sender
FROM events
WHERE (type = 'm.room.encrypted' OR type = 'm.room.message')
AND origin_server_ts >= DATE_PART('epoch', NOW() - INTERVAL '1 month') * 1000
GROUP BY sender
ORDER BY COUNT(*) DESC
LIMIT 20;
SELECT e.room_id, r.name, e.event_id, e.type, e.content, j.json
FROM events e
LEFT JOIN event_json j USING (room_id)
LEFT JOIN room_stats_state r USING (room_id)
WHERE sender = '@LOGIN:example.com'
AND e.type = 'm.room.message'
ORDER BY stream_ordering DESC
LIMIT 100;
Documentation for psql command line parameters: https://www.postgresql.org/docs/current/app-psql.html
Sort and order with SQL
SELECT COUNT(*), event_json.room_id, room_stats_state.name
FROM event_json, room_stats_state
WHERE room_stats_state.room_id = event_json.room_id
GROUP BY event_json.room_id, room_stats_state.name
ORDER BY COUNT(*) DESC
LIMIT 50;
You get the same information when you use the
admin API.
SELECT rss.room_id, rss.name, rss.canonical_alias, rss.topic, rss.encryption,
rsc.joined_members, rsc.local_users_in_room, rss.join_rules
FROM room_stats_state rss
LEFT JOIN room_stats_current rsc USING (room_id)
WHERE room_id IN (
'!OGEhHVWSdvArJzumhm:matrix.org',
'!YTvKGNlinIzlkMTVRl:matrix.org'
);
As a general rule, I encourage people who want to understand the deepest darkest secrets of the database schema to drop by #synapse-dev:matrix.org and ask questions.
However, one question that comes up frequently is that of how "state groups" work, and why the state_groups_state table gets so big, so here's an attempt to answer that question.
We need to be able to relatively quickly calculate the state of a room at any point in that room's history. In other words, we need to know the state of the room at each event in that room. This is done as follows:
A sequence of events where the state is the same are grouped together into a state_group; the mapping is recorded in event_to_state_groups. (Technically speaking, since a state event usually changes the state in the room, we are recording the state of the room after the given event id: which is to say, to a handwavey simplification, the first event in a state group is normally a state event, and others in the same state group are normally non-state-events.)
state_groups records, for each state group, the id of the room that we're looking at, and also the id of the first event in that group. (I'm not sure if that event id is used much in practice.)
Now, if we stored all the room state for each state_group, that would be a huge amount of data. Instead, for each state group, we normally store the difference between the state in that group and some other state group, and only occasionally (every 100 state changes or so) record the full state.
So, most state groups have an entry in state_group_edges (don't ask me why it's not a column in state_groups) which records the previous state group in the room, and state_groups_state records the differences in state since that previous state group.
A full state group just records the event id for each piece of state in the room at that point.
There are various reasons that we can end up creating many more state groups than we need: see https://github.com/matrix-org/synapse/issues/3364 for more details.
There is a tool at https://github.com/matrix-org/rust-synapse-compress-state which can compress the state_groups_state on a room by-room basis (essentially, it reduces the number of "full" state groups). This can result in dramatic reductions of the storage used.
Logger name (synapse.access.(http\|https).<tag>, where 'tag' is defined in the listeners config section, normally the port)
CCCC
Line number in code
DDDD
Log Level
EEEE
Request Identifier (This identifier is shared by related log lines)
FFFF
Source IP (Or X-Forwarded-For if enabled)
GGGG
Server Port
HHHH
Federated Server or Local User making request (blank if unauthenticated or not supplied). If this is of the form `@aaa:example.com
IIII
Total Time to process the request
JJJJ
Time to send response over network once generated (this may be negative if the socket is closed before the response is generated)
KKKK
Userland CPU time
LLLL
System CPU time
MMMM
Total time waiting for a free DB connection from the pool across all parallel DB work from this request
NNNN
Total time waiting for response to DB queries across all parallel DB work from this request
OOOO
Count of DB transactions performed
PPPP
Response body size
QQQQ
Response status code Suffixed with ! if the socket was closed before the response was generated. A 499! status code indicates that Synapse also cancelled request processing after the socket was closed.
RRRR
Request
SSSS
User-agent
TTTT
Events fetched from DB to service this request (note that this does not include events fetched from the cache)
MMMM / NNNN can be greater than IIII if there are multiple slow database queries
running in parallel.
Some actions can result in multiple identical http requests, which will return
the same data, but only the first request will report time/transactions in
KKKK/LLLL/MMMM/NNNN/OOOO - the others will be awaiting the first query to return a
response and will simultaneously return with the first request, but with very
small processing times.
If your server already has an admin account you should use the
User Admin API
to promote other accounts to become admins.
If you don't have any admin accounts yet you won't be able to use the admin API,
so you'll have to edit the database manually. Manually editing the database is
generally not recommended so once you have an admin account: use the admin APIs
to make further changes.
UPDATE users SET admin = 1 WHERE name = '@foo:bar.com';
Synapse includes a Python command to export data for a specific user. It takes the homeserver
configuration file and the full Matrix ID of the user to export:
The directory to store the export data in can be customised with the
--output-directory parameter; ensure that the provided directory is
empty. If this parameter is not provided, Synapse defaults to creating
a temporary directory (which starts with "synapse-exfiltrate") in /tmp,
/var/tmp, or /usr/tmp, in that order.
The media_ids folder contains only the metadata of the media uploaded by the user.
It does not contain the media itself.
Furthermore, only the media_ids that Synapse manages itself are exported.
If another media repository (e.g. matrix-media-repo)
is used, the data must be exported separately.
With the media_ids the media files can be downloaded.
Media that have been sent in encrypted rooms are only retrieved in encrypted form.
The following script can help with download the media files:
#!/usr/bin/env bash
# Parameters
#
# source_directory: Directory which contains the export with the media_ids.
# target_directory: Directory into which all files are to be downloaded.
# repository_url: Address of the media repository resp. media worker.
# serverName: Name of the server (`server_name` from homeserver.yaml).
#
# Example:
# ./download_media.sh /tmp/export_data/media_ids/ /tmp/export_data/media_files/ http://localhost:8008 matrix.example.com
source_directory=$1
target_directory=$2
repository_url=$3
serverName=$4
mkdir -p $target_directory
for file in $source_directory/*; do
filename=$(basename ${file})
url=$repository_url/_matrix/media/v3/download/$serverName/$filename
echo "Downloading $filename - $url"
if ! wget -o /dev/null -P $target_directory $url; then
echo "Could not download $filename"
fi
done
Deleting your database is unlikely to make anything better.
It's easy to make the mistake of thinking that you can start again from a clean
slate by dropping your database, but things don't work like that in a federated
network: lots of other servers have information about your server.
For example: other servers might think that you are in a room, your server will
think that you are not, and you'll probably be unable to interact with that room
in a sensible way ever again.
In general, there are better solutions to any problem than dropping the database.
Come and seek help in https://matrix.to/#/#synapse:matrix.org.
There are two exceptions when it might be sensible to delete your database and start again:
You have never joined any rooms which are federated with other servers. For
instance, a local deployment which the outside world can't talk to.
You are changing the server_name in the homeserver configuration. In effect
this makes your server a completely new one from the point of view of the network,
so in this case it makes sense to start with a clean database.
(In both cases you probably also want to clear out the media_store.)
Synapse tags each log line according to the HTTP request it is processing. When
it finishes processing each request, it logs a line containing the words
Processed request: . For example:
Here we can see that the request has been tagged with GET-37. (The tag depends
on the method of the HTTP request, so might start with GET-, PUT-, POST-,
OPTIONS- or DELETE-.) So to find all lines corresponding to this request, we can do:
grep 'GET-37' homeserver.log
If you want to paste that output into a github issue or matrix room, please
remember to surround it with triple-backticks (```) to make it legible
(see quoting code).
SELECT s.canonical_alias, g.room_id, count(*) AS num_rows
FROM
state_groups_state AS g,
room_stats_state AS s
WHERE g.room_id = s.room_id
GROUP BY s.canonical_alias, g.room_id
ORDER BY num_rows desc
LIMIT 10;
You can also use the List Room API
and order_bystate_events.
The typical failure mode here is that you send an invitation to someone
to join a room or direct chat, but when they go to accept it, they get an
error (typically along the lines of "Invalid signature"). They might see
something like the following in their logs:
2019-09-11 19:32:04,271 - synapse.federation.transport.server - 288 - WARNING - GET-11752 - authenticate_request failed: 401: Invalid signature for server <server> with key ed25519:a_EqML: Unable to verify signature for <server>
This is normally caused by a misconfiguration in your reverse-proxy. See the reverse proxy docs and double-check that your settings are correct.
First, ensure you are running the latest version of Synapse, using Python 3
with a PostgreSQL database.
Synapse's architecture is quite RAM hungry currently - we deliberately
cache a lot of recent room data and metadata in RAM in order to speed up
common requests. We'll improve this in the future, but for now the easiest
way to either reduce the RAM usage (at the risk of slowing things down)
is to set the almost-undocumented SYNAPSE_CACHE_FACTOR environment
variable. The default is 0.5, which can be decreased to reduce RAM usage
in memory constrained environments, or increased if performance starts to
degrade.
However, degraded performance due to a low cache factor, common on
machines with slow disks, often leads to explosions in memory use due
backlogged requests. In this case, reducing the cache factor will make
things worse. Instead, try increasing it drastically. 2.0 is a good
starting value.
Using libjemalloc can also yield a significant
improvement in overall memory use, and especially in terms of giving back
RAM to the OS. To use it, the library must simply be put in the
LD_PRELOAD environment variable when launching Synapse. On Debian, this
can be done by installing the libjemalloc1 package and adding this
line to /etc/default/matrix-synapse:
This made a significant difference on Python 2.7 - it's unclear how
much of an improvement it provides on Python 3.x.
If you're encountering high CPU use by the Synapse process itself, you
may be affected by a bug with presence tracking that leads to a
massive excess of outgoing federation requests (see discussion). If metrics
indicate that your server is also issuing far more outgoing federation
requests than can be accounted for by your users' activity, this is a
likely cause. The misbehavior can be worked around by disabling presence
in the Synapse config file: see here.
If Synapse runs out of file handles, it typically fails badly - live-locking
at 100% CPU, and/or failing to accept new TCP connections (blocking the
connecting client). Matrix currently can legitimately use a lot of file handles,
thanks to busy rooms like #matrix:matrix.org containing hundreds of participating
servers. The first time a server talks in a room it will try to connect
simultaneously to all participating servers, which could exhaust the available
file descriptors between DNS queries & HTTPS sockets, especially if DNS is slow
to respond. (We need to improve the routing algorithm used to be better than
full mesh, but as of March 2019 this hasn't happened yet).
If you hit this failure mode, we recommend increasing the maximum number of
open file handles to be at least 4096 (assuming a default of 1024 or 256).
This is typically done by editing /etc/security/limits.conf
Separately, Synapse may leak file handles if inbound HTTP requests get stuck
during processing - e.g. blocked behind a lock or talking to a remote server etc.
This is best diagnosed by matching up the 'Received request' and 'Processed request'
log lines and looking for any 'Processed request' lines which take more than
a few seconds to execute. Please let us know at #synapse:matrix.org if
you see this failure mode so we can help debug it, however.
Everyone is welcome to contribute code to matrix.org
projects, provided that they are willing to
license their contributions under the same license as the project itself. We
follow a simple 'inbound=outbound' model for contributions: the act of
submitting an 'inbound' contribution means that the contributor agrees to
license the code under the same terms as the project's overall 'outbound'
license - in our case, this is almost always Apache Software License v2 (see
LICENSE).
If you are running Windows, the Windows Subsystem for Linux (WSL) is strongly
recommended for development. More information about WSL can be found at
https://docs.microsoft.com/en-us/windows/wsl/install. Running Synapse natively
on Windows is not officially supported.
The code of Synapse is written in Python 3. To do pretty much anything, you'll need a recent version of Python 3. Your Python also needs support for virtual environments. This is usually built-in, but some Linux distributions like Debian and Ubuntu split it out into its own package. Running sudo apt install python3-venv should be enough.
A recent version of the Rust compiler is needed to build the native modules. The
easiest way of installing the latest version is to use rustup.
Synapse can connect to PostgreSQL via the psycopg2 Python library. Building this library from source requires access to PostgreSQL's C header files. On Debian or Ubuntu Linux, these can be installed with sudo apt install libpq-dev.
Synapse has an optional, improved user search with better Unicode support. For that you need the development package of libicu. On Debian or Ubuntu Linux, this can be installed with sudo apt install libicu-dev.
The preferred and easiest way to contribute changes is to fork the relevant
project on GitHub, and then create a pull request to ask us to pull your
changes into our repo.
Before installing the Python dependencies, make sure you have installed a recent version
of Rust (see the "What do I need?" section above). The easiest way of installing the
latest version is to use rustup.
Synapse uses the poetry project to manage its dependencies
and development environment. Once you have installed Python 3 and added the
source, you should install poetry.
Of their installation methods, we recommend
installing poetry using pipx,
pip install --user pipx
pipx install poetry==1.5.1 # Problems with Poetry 1.6, see https://github.com/matrix-org/synapse/issues/16147
There is a growing amount of documentation located in the
docs
directory, with a rendered version available online.
This documentation is intended primarily for sysadmins running their
own Synapse instance, as well as developers interacting externally with
Synapse.
docs/development
exists primarily to house documentation for
Synapse developers.
docs/admin_api houses documentation
regarding Synapse's Admin API, which is used mostly by sysadmins and external
service developers.
We welcome improvements and additions to our documentation itself! When
writing new pages, please
build docs to a book
to check that your contributions render correctly. The docs are written in
GitHub-Flavoured Markdown.
Some documentation also exists in Synapse's GitHub
Wiki, although this is primarily
contributed to by community authors.
When changes are made to any Rust code then you must call either poetry install
or maturin develop (if installed) to rebuild the Rust code. Using maturin
is quicker than poetry install, so is recommended when making frequent
changes to the Rust code.
The unit tests run parts of Synapse, including your changes, to see if anything
was broken. They are slower than the linters but will typically catch more errors.
poetry run trial tests
You can run unit tests in parallel by specifying -jX argument to trial where X is the number of parallel runners you want. To use 4 cpu cores, you would run them like:
poetry run trial -j4 tests
If you wish to only run some unit tests, you may specify
another module instead of tests - or a test class or a method:
poetry run trial tests.rest.admin.test_room tests.handlers.test_admin.ExfiltrateData.test_invite
If your tests fail, you may wish to look at the logs (the default log level is ERROR):
less _trial_temp/test.log
To increase the log level for the tests, set SYNAPSE_TEST_LOG_LEVEL:
SYNAPSE_TEST_LOG_LEVEL=DEBUG poetry run trial tests
By default, tests will use an in-memory SQLite database for test data. For additional
help with debugging, one can use an on-disk SQLite database file instead, in order to
review database state during and after running tests. This can be done by setting
the SYNAPSE_TEST_PERSIST_SQLITE_DB environment variable. Doing so will cause the
database state to be stored in a file named test.db under the trial process'
working directory. Typically, this ends up being _trial_temp/test.db. For example:
SYNAPSE_TEST_PERSIST_SQLITE_DB=1 poetry run trial tests
The database file can then be inspected with:
sqlite3 _trial_temp/test.db
Note that the database file is cleared at the beginning of each test run. Thus it
will always only contain the data generated by the last run test. Though generally
when debugging, one is only running a single test anyway.
Invoking trial as above will use an in-memory SQLite database. This is great for
quick development and testing. However, we recommend using a PostgreSQL database
in production (and indeed, we have some code paths specific to each database).
This means that we need to run our unit tests against PostgreSQL too. Our CI does
this automatically for pull requests and release candidates, but it's sometimes
useful to reproduce this locally.
You don't need to specify the host, user, port or password if your Postgres
server is set to authenticate you over the UNIX socket (i.e. if the psql command
works without further arguments).
Your Postgres account needs to be able to create databases; see the postgres
docs for ALTER ROLE.
The integration tests are a more comprehensive suite of tests. They
run a full version of Synapse, including your changes, to check if
anything was broken. They are slower than the unit tests but will
typically catch more errors.
The following command will let you run the integration test with the most common
configuration:
$ docker run --rm -it -v /path/where/you/have/cloned/the/repository\:/src:ro -v /path/to/where/you/want/logs\:/logs matrixdotorg/sytest-synapse:focal
(Note that the paths must be full paths! You could also write $(realpath relative/path) if needed.)
This configuration should generally cover your needs.
To run with Postgres, supply the -e POSTGRES=1 -e MULTI_POSTGRES=1 environment flags.
To run with Synapse in worker mode, supply the -e WORKERS=1 -e REDIS=1 environment flags (in addition to the Postgres flags).
Complement is a suite of black box tests that can be run on any homeserver implementation. It can also be thought of as end-to-end (e2e) tests.
It's often nice to develop on Synapse and write Complement tests at the same time.
Here is how to run your local Synapse checkout against your local Complement checkout.
(checkout complement alongside your synapse checkout)
To run a specific test file, you can pass the test name at the end of the command. The name passed comes from the naming structure in your Complement tests. If you're unsure of the name, you can do a full run and copy it from the test output:
The above will run a monolithic (single-process) Synapse with SQLite as the database. For other configurations, try:
Passing POSTGRES=1 as an environment variable to use the Postgres database instead.
Passing WORKERS=1 as an environment variable to use a workerised setup instead. This option implies the use of Postgres.
If setting WORKERS=1, optionally set WORKER_TYPES= to declare which worker
types you wish to test. A simple comma-delimited string containing the worker types
defined from the WORKERS_CONFIG template in
here.
A safe example would be WORKER_TYPES="federation_inbound, federation_sender, synchrotron".
See the worker documentation for additional information on workers.
Passing ASYNCIO_REACTOR=1 as an environment variable to use the Twisted asyncio reactor instead of the default one.
Passing PODMAN=1 will use the podman container runtime, instead of docker.
Passing UNIX_SOCKETS=1 will utilise Unix socket functionality for Synapse, Redis, and Postgres(when applicable).
To increase the log level for the tests, set SYNAPSE_TEST_LOG_LEVEL, e.g:
If you're curious what the database looks like after you run some tests, here are some steps to get you going in Synapse:
In your Complement test comment out defer deployment.Destroy(t) and replace with defer time.Sleep(2 * time.Hour) to keep the homeserver running after the tests complete
Start the Complement tests
Find the name of the container, docker ps -f name=complement_ (this will filter for just the Compelement related Docker containers)
Access the container replacing the name with what you found in the previous step: docker exec -it complement_1_hs_with_application_service.hs1_2 /bin/bash
All changes, even minor ones, need a corresponding changelog / newsfragment
entry. These are managed by Towncrier.
To create a changelog entry, make a new file in the changelog.d directory named
in the format of PRnumber.type. The type can be one of the following:
feature
bugfix
docker (for updates to the Docker image)
doc (for updates to the documentation)
removal (also used for deprecations)
misc (for internal-only changes)
This file will become part of our changelog at the next
release, so the content of the file should be a short description of your
change in the same style as the rest of the changelog. The file can contain Markdown
formatting, and must end with a full stop (.) or an exclamation mark (!) for
consistency.
Adding credits to the changelog is encouraged, we value your
contributions and would like to have you shouted out in the release notes!
For example, a fix in PR #1234 would have its changelog entry in
changelog.d/1234.bugfix, and contain content like:
The security levels of Florbs are now validated when received
via the /federation/florb endpoint. Contributed by Jane Matrix.
If there are multiple pull requests involved in a single bugfix/feature/etc,
then the content for each changelog.d file should be the same. Towncrier will
merge the matching files together into a single changelog entry when we come to
release.
Obviously, you don't know if you should call your newsfile
1234.bugfix or 5678.bugfix until you create the PR, which leads to a
chicken-and-egg problem.
There are two options for solving this:
Open the PR without a changelog file, see what number you got, and then
add the changelog file to your branch, or:
Look at the list of all
issues/PRs, add one to the
highest number you see, and quickly open the PR before somebody else claims
your number.
This
script
might be helpful if you find yourself doing this a lot.
Sorry, we know it's a bit fiddly, but it's really helpful for us when we come
to put together a release!
Changes which affect the debian packaging files (in debian) are an
exception to the rule that all changes require a changelog.d file.
In this case, you will need to add an entry to the debian changelog for the
next release. For this, run the following command:
dch
This will make up a new version number (if there isn't already an unreleased
version in flight), and open an editor where you can add a new changelog entry.
(Our release process will ensure that the version number and maintainer name is
corrected for the release.)
If your change affects both the debian packaging and files outside the debian
directory, you will need both a regular newsfragment and an entry in the
debian changelog. (Though typically such changes should be submitted as two
separate pull requests.)
In order to have a concrete record that your contribution is intentional
and you agree to license it under the same terms as the project's license, we've adopted the
same lightweight approach that the Linux Kernel
submitting patches process,
Docker, and many other
projects use: the DCO (Developer Certificate of Origin).
This is a simple declaration that you wrote
the contribution or otherwise have the right to contribute it to Matrix:
Developer Certificate of Origin
Version 1.1
Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
660 York Street, Suite 102,
San Francisco, CA 94110 USA
Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.
Developer's Certificate of Origin 1.1
By making a contribution to this project, I certify that:
(a) The contribution was created in whole or in part by me and I
have the right to submit it under the open source license
indicated in the file; or
(b) The contribution is based upon previous work that, to the best
of my knowledge, is covered under an appropriate open source
license and I have the right under that license to submit that
work with modifications, whether created in whole or in part
by me, under the same open source license (unless I am
permitted to submit under a different license), as indicated
in the file; or
(c) The contribution was provided directly to me by some other
person who certified (a), (b) or (c) and I have not modified
it.
(d) I understand and agree that this project and the contribution
are public and that a record of the contribution (including all
personal information I submit with it, including my sign-off) is
maintained indefinitely and may be redistributed consistent with
this project or the open source license(s) involved.
If you agree to this for your contribution, then all that's needed is to
include the line in your commit or pull request comment:
Signed-off-by: Your Name <your@email.example.org>
We accept contributions under a legally identifiable name, such as
your name on government documentation or common-law names (names
claimed by legitimate usage or repute). Unfortunately, we cannot
accept anonymous contributions at this time.
Git allows you to add this signoff automatically when using the -s
flag to git commit, which uses the name and email set in your
user.name and user.email git configs.
If you would like to provide your legal name privately to the Matrix.org
Foundation (instead of in a public commit or comment), you can do so
by emailing your legal name and a link to the pull request to
dco@matrix.org.
It helps to include "sign off" or similar in the subject line. You will then
be instructed further.
Once private sign off is complete, doing so for future contributions will not
be required.
Once the Pull Request is opened, you will see a few things:
our automated CI (Continuous Integration) pipeline will run (again) the linters, the unit tests, the integration tests and more;
one or more of the developers will take a look at your Pull Request and offer feedback.
From this point, you should:
Look at the results of the CI pipeline.
If there is any error, fix the error.
If a developer has requested changes, make these changes and let us know if it is ready for a developer to review again.
A pull request is a conversation, if you disagree with the suggestions, please respond and discuss it.
Create a new commit with the changes.
Please do NOT overwrite the history. New commits make the reviewer's life easier.
Push this commits to your Pull Request.
Back to 1.
Once the pull request is ready for review again please re-request review from whichever developer did your initial
review (or leave a comment in the pull request that you believe all required changes have been done).
Once both the CI and the developers are happy, the patch will be merged into Synapse and released shortly!
That's it! Matrix is a very open and collaborative project as you might expect
given our obsession with open communication. If we're going to successfully
matrix together all the fragmented communication technologies out there we are
reliant on contributions and collaboration from the community to do so. So
please get involved - and we hope you have as much fun hacking on Matrix as we
do!
The Synapse codebase uses a number of code formatting tools in order to
quickly and automatically check for formatting (and sometimes logical)
errors in code.
See the contributing guide for instructions
on how to install the above tools and run the linters.
It's worth noting that modern IDEs and text editors can run these tools
automatically on save. It may be worth looking into whether this
functionality is supported in your editor for a more convenient
development workflow. It is not, however, recommended to run mypy
on save as it takes a while and can be very resource intensive.
When adding a configuration option to the code, if several settings are grouped into a single dict, ensure that your code
correctly handles the top-level option being set to None (as it will be if no sub-options are enabled).
The configuration manual acts as a
reference to Synapse's configuration options for server administrators.
Remember that many readers will be unfamiliar with YAML and server
administration in general, so it is important that when you add
a configuration option the documentation be as easy to understand as possible, which
includes following a consistent format.
Some guidelines follow:
Each option should be listed in the config manual with the following format:
The name of the option, prefixed by ###.
A comment which describes the default behaviour (i.e. what
happens if the setting is omitted), as well as what the effect
will be if the setting is changed.
An example setting, using backticks to define the code block
For boolean (on/off) options, convention is that this example
should be the opposite to the default. For other options, the example should give
some non-default value which is likely to be useful to the reader.
There should be a horizontal rule between each option, which can be achieved by adding --- before and
after the option.
true and false are spelt thus (as opposed to True, etc.)
Use the module sub-option to add a module under modules to extend functionality.
The module setting then has a sub-option, config, which can be used to define some configuration
for the module.
Note that the sample configuration is generated from the synapse code
and is maintained by a script, scripts-dev/generate_sample_config.sh.
Making sure that the output from this script matches the desired format
is left as an exercise for the reader!
The Synapse team works off a shared review queue -- any new pull requests for
Synapse (or related projects) has a review requested from the entire team. Team
members should process this queue using the following rules:
Any high urgency pull requests (e.g. fixes for broken continuous integration
or fixes for release blockers);
Follow-up reviews for pull requests which have previously received reviews;
Any remaining pull requests.
For the latter two categories above, older pull requests should be prioritised.
It is explicit that there is no priority given to pull requests from the team
(vs from the community). If a pull request requires a quick turn around, please
explicitly communicate this via #synapse-dev:matrix.org
or as a comment on the pull request.
Once an initial review has been completed and the author has made additional changes,
follow-up reviews should go back to the same reviewer. This helps build a shared
context and conversation between author and reviewer.
As a team we aim to keep the number of inflight pull requests to a minimum to ensure
that ongoing work is finished before starting new work.
To communicate to the rest of the team the status of each pull request, team
members should do the following:
Assign themselves to the pull request (they should be left assigned to the
pull request until it is merged, closed, or are no longer the reviewer);
Review the pull request by leaving comments, questions, and suggestions;
Mark the pull request appropriately (as needing changes or accepted).
If you are unsure about a particular part of the pull request (or are not confident
in your understanding of part of the code) then ask questions or request review
from the team again. When requesting review from the team be sure to leave a comment
with the rationale on why you're putting it back in the queue.
If a bug is found after release that is deemed severe enough (by a combination
of the impacted users and the impact on those users) then a bugfix release may
be issued. This may be at any point in the release cycle.
Security will sometimes be backported to the previous version and released
immediately before the next release candidate. An example of this might be:
Day 0: Synapse N - 1 is released.
Day 7: Synapse (N - 1).1 is released as Synapse N - 1 + the security fix.
Day 7: Synapse N release candidate 1 is released (including the security fix).
Depending on the impact and complexity of security fixes, multiple fixes might
be held to be released together.
In some cases, a pre-disclosure of a security release will be issued as a notice
to Synapse operators that there is an upcoming security release. These can be
found in the security category of the Matrix blog.
In an ideal world, our git commit history would be a linear progression of
commits each of which contains a single change building on what came
before. Here, by way of an arbitrary example, is the top of git log --graph b2dba0607:
Note how the commit comment explains clearly what is changing and why. Also
note the absence of merge commits, as well as the absence of commits called
things like (to pick a few culprits):
“pep8”, “fix broken
test”,
“oops”,
“typo”, or “Who's
the president?”.
There are a number of reasons why keeping a clean commit history is a good
thing:
From time to time, after a change lands, it turns out to be necessary to
revert it, or to backport it to a release branch. Those operations are
much easier when the change is contained in a single commit.
Similarly, it's much easier to answer questions like “is the fix for
/publicRooms on the release branch?” if that change consists of a single
commit.
Likewise: “what has changed on this branch in the last week?” is much
clearer without merges and “pep8” commits everywhere.
Sometimes we need to figure out where a bug got introduced, or some
behaviour changed. One way of doing that is with git bisect: pick an
arbitrary commit between the known good point and the known bad point, and
see how the code behaves. However, that strategy fails if the commit you
chose is the middle of someone's epic branch in which they broke the world
before putting it back together again.
One counterargument is that it is sometimes useful to see how a PR evolved as
it went through review cycles. This is true, but that information is always
available via the GitHub UI (or via the little-known refs/pull
namespace).
Of course, in reality, things are more complicated than that. We have release
branches as well as develop and master, and we deliberately merge changes
between them. Bugs often slip through and have to be fixed later. That's all
fine: this not a cast-iron rule which must be obeyed, but an ideal to aim
towards.
Ok, so that's what we'd like to achieve. How do we achieve it?
The TL;DR is: when you come to merge a pull request, you probably want to
“squash and merge”:
.
(This applies whether you are merging your own PR, or that of another
contributor.)
“Squash and merge”1 takes all of the changes in the
PR, and bundles them into a single commit. GitHub gives you the opportunity to
edit the commit message before you confirm, and normally you should do so,
because the default will be useless (again: * woops typo is not a useful
thing to keep in the historical record).
The main problem with this approach comes when you have a series of pull
requests which build on top of one another: as soon as you squash-merge the
first PR, you'll end up with a stack of conflicts to resolve in all of the
others. In general, it's best to avoid this situation in the first place by
trying not to have multiple related PRs in flight at the same time. Still,
sometimes that's not possible and doing a regular merge is the lesser evil.
Ultimately: this is not a hard-and-fast-rule. If in doubt, ask yourself “do
each of the commits I am about to merge make sense in their own right”, but
remember that we're just doing our best to balance “keeping the commit history
clean” with other factors.
A lotofwords have been
written in the past about git branching models (no really, a
lot). I tend to
think the whole thing is overblown. Fundamentally, it's not that
complicated. Here's how we do it.
Let's start with a picture:
It looks complicated, but it's really not. There's one basic rule: anyone is
free to merge from any more-stable branch to any less-stable branch at
any time2. (The principle behind this is that if a
change is good enough for the more-stable branch, then it's also good enough go
put in a less-stable branch.)
Meanwhile, merging (or squashing, as per the above) from a less-stable to a
more-stable branch is a deliberate action in which you want to publish a change
or a set of changes to (some subset of) the world: for example, this happens
when a PR is landed, or as part of our release process.
So, what counts as a more- or less-stable branch? A little reflection will show
that our active branches are ordered thus, from more-stable to less-stable:
master (tracks our last release).
release-vX.Y (the branch where we prepare the next release)3.
The corollary is: if you have a bugfix that needs to land in both
release-vX.Yanddevelop, then you should base your PR on
release-vX.Y, get it merged there, and then merge from release-vX.Y to
develop. (If a fix lands in develop and we later need it in a
release-branch, we can of course cherry-pick it, but landing it in the release
branch first helps reduce the chance of annoying conflicts.)
[1]: “Squash and merge” is GitHub's term for this
operation. Given that there is no merge involved, I'm not convinced it's the
most intuitive name. ^
[3]: Very, very occasionally (I think this has happened once in
the history of Synapse), we've had two releases in flight at once. Obviously,
release-v1.2 is more-stable than release-v1.3. ^
The demo setup allows running three federation Synapse servers, with server
names localhost:8480, localhost:8481, and localhost:8482.
You can access them via any Matrix client over HTTP at localhost:8080,
localhost:8081, and localhost:8082 or over HTTPS at localhost:8480,
localhost:8481, and localhost:8482.
To enable the servers to communicate, self-signed SSL certificates are generated
and the servers are configured in a highly insecure way, including:
Not checking certificates over federation.
Not verifying keys.
The servers are configured to store their data under demo/8080, demo/8081, and
demo/8082. This includes configuration, logs, SQLite databases, and media.
Note that when joining a public room on a different homeserver via "#foo:bar.net",
then you are (in the current implementation) joining a room with room_id "foo".
This means that it won't work if your homeserver already has a room with that
name.
OpenTracing is a semi-standard being adopted by a number of distributed
tracing platforms. It is a common api for facilitating vendor-agnostic
tracing instrumentation. That is, we can use the OpenTracing api and
select one of a number of tracer implementations to do the heavy lifting
in the background. Our current selected implementation is Jaeger.
OpenTracing is a tool which gives an insight into the causal
relationship of work done in and between servers. The servers each track
events and report them to a centralised server - in Synapse's case:
Jaeger. The basic unit used to represent events is the span. The span
roughly represents a single piece of work that was done and the time at
which it occurred. A span can have child spans, meaning that the work of
the child had to be completed for the parent span to complete, or it can
have follow-on spans which represent work that is undertaken as a result
of the parent but is not depended on by the parent to in order to
finish.
Since this is undertaken in a distributed environment a request to
another server, such as an RPC or a simple GET, can be considered a span
(a unit or work) for the local server. This causal link is what
OpenTracing aims to capture and visualise. In order to do this metadata
about the local server's span, i.e the 'span context', needs to be
included with the request to the remote.
It is up to the remote server to decide what it does with the spans it
creates. This is called the sampling policy and it can be configured
through Jaeger's settings.
By default, Synapse will publish traces to Jaeger on localhost.
If Jaeger is hosted elsewhere, point Synapse to the correct host by setting
opentracing.jaeger_config.local_agent.reporting_hostin the Synapse configuration
or by setting the JAEGER_AGENT_HOST environment variable to the desired address.
Latest documentation is probably at
https://www.jaegertracing.io/docs/latest/getting-started.
OpenTracing is not enabled by default. It must be enabled in the
homeserver config by adding the opentracing option to your config file. You can find
documentation about how to do this in the config manual under the header 'Opentracing'.
See below for an example Opentracing configuration:
The homeserver whitelist is configured using regular expressions. A list
of regular expressions can be given and their union will be compared
when propagating any spans contexts to another homeserver.
Though it's mostly safe to send and receive span contexts to and from
untrusted users since span contexts are usually opaque ids it can lead
to two problems, namely:
If the span context is marked as sampled by the sending homeserver
the receiver will sample it. Therefore two homeservers with wildly
different sampling policies could incur higher sampling counts than
intended.
Sending servers can attach arbitrary data to spans, known as
'baggage'. For safety this has been disabled in Synapse but that
doesn't prevent another server sending you baggage which will be
logged to OpenTracing's logs.
Synapse supports splitting its datastore across multiple physical databases (which can
be useful for large installations), and the schema files are therefore split according
to the logical database they apply to.
At the time of writing, the following "logical" databases are supported:
state - used to store Matrix room state (more specifically, state_groups,
their relationships and contents).
main - stores everything else.
Additionally, the common directory contains schema files for tables which must be
present on all physical databases.
Synapse manages its database schema via "schema versions". These are mainly used to
help avoid confusion if the Synapse codebase is rolled back after the database is
updated. They work as follows:
The Synapse codebase defines a constant synapse.storage.schema.SCHEMA_VERSION
which represents the expectations made about the database by that version. For
example, as of Synapse v1.36, this is 59.
The database stores a "compatibility version" in
schema_compat_version.compat_version which defines the SCHEMA_VERSION of the
oldest version of Synapse which will work with the database. On startup, if
compat_version is found to be newer than SCHEMA_VERSION, Synapse will refuse to
start.
Synapse automatically updates this field from
synapse.storage.schema.SCHEMA_COMPAT_VERSION.
Whenever a backwards-incompatible change is made to the database format (normally
via a delta file), synapse.storage.schema.SCHEMA_COMPAT_VERSION is also updated
so that administrators can not accidentally roll back to a too-old version of Synapse.
Generally, the goal is to maintain compatibility with at least one or two previous
releases of Synapse, so any substantial change tends to require multiple releases and a
bit of forward-planning to get right.
As a worked example: we want to remove the room_stats_historical table. Here is how it
might pan out.
Replace any code that reads from room_stats_historical with alternative
implementations, but keep writing to it in case of rollback to an earlier version.
Also, increase synapse.storage.schema.SCHEMA_VERSION. In this
instance, there is no existing code which reads from room_stats_historical, so
our starting point is:
Next (say in Synapse v1.37.0): remove the code that writes to
room_stats_historical, but don’t yet remove the table in case of rollback to
v1.36.0. Again, we increase synapse.storage.schema.SCHEMA_VERSION, but
because we have not broken compatibility with v1.36, we do not yet update
SCHEMA_COMPAT_VERSION. We now have:
Later (say in Synapse v1.38.0): we can remove the table altogether. This will
break compatibility with v1.36.0, so we must update SCHEMA_COMPAT_VERSION accordingly.
There is no need to update synapse.storage.schema.SCHEMA_VERSION, since there is no
change to the Synapse codebase here. So we end up with:
In the full_schemas directories, only the most recently-numbered snapshot is used
(54 at the time of writing). Older snapshots (eg, 16) are present for historical
reference only.
Delta files define the steps required to upgrade the database from an earlier version.
They can be written as either a file containing a series of SQL statements, or a Python
module.
Synapse remembers which delta files it has applied to a database (they are stored in the
applied_schema_deltas table) and will not re-apply them (even if a given file is
subsequently updated).
Delta files should be placed in a directory named synapse/storage/schema/<database>/delta/<version>/.
They are applied in alphanumeric order, so by convention the first two characters
of the filename should be an integer such as 01, to put the file in the right order.
These should be named *.sql, or — for changes which should only be applied for a
given database engine — *.sql.posgres or *.sql.sqlite. For example, a delta which
adds a new column to the foo table might be called 01add_bar_to_foo.sql.
Note that our SQL parser is a bit simple - it understands comments (-- and /*...*/),
but complex statements which require a ; in the middle of them (such as CREATE TRIGGER) are beyond it and you'll have to use a Python delta file.
For more flexibility, a delta file can take the form of a python module. These should
be named *.py. Note that database-engine-specific modules are not supported here –
instead you can write if isinstance(database_engine, PostgresEngine) or similar.
A Python delta module should define either or both of the following functions:
import synapse.config.homeserver
import synapse.storage.engines
import synapse.storage.types
def run_create(
cur: synapse.storage.types.Cursor,
database_engine: synapse.storage.engines.BaseDatabaseEngine,
) -> None:
"""Called whenever an existing or new database is to be upgraded"""
...
def run_upgrade(
cur: synapse.storage.types.Cursor,
database_engine: synapse.storage.engines.BaseDatabaseEngine,
config: synapse.config.homeserver.HomeServerConfig,
) -> None:
"""Called whenever an existing database is to be upgraded."""
...
It is sometimes appropriate to perform database migrations as part of a background
process (instead of blocking Synapse until the migration is done). In particular,
this is useful for migrating data when adding new columns or tables.
Pending background updates stored in the background_updates table and are denoted
by a unique name, the current status (stored in JSON), and some dependency information:
Whether the update requires a previous update to be complete.
A rough ordering for which to complete updates.
A new background updates needs to be added to the background_updates table:
There are a few types of updates that can be performed, see the BackgroundUpdater:
register_background_update_handler: A generic handler for custom SQL
register_background_index_update: Create an index in the background
register_background_validate_constraint: Validate a constraint in the background
(PostgreSQL-only)
register_background_validate_constraint_and_delete_rows: Similar to
register_background_validate_constraint, but deletes rows which don't fit
the constraint.
For register_background_update_handler, the generic handler must track progress
and then finalize the background update:
Synapse will attempt to rate-limit how often background updates are run via the
given batch-size and the returned number of processed entries (and how long the
function took to run). See
background update controller callbacks.
Boolean columns require special treatment, since SQLite treats booleans the
same as integers.
Any new boolean column must be added to the BOOLEAN_COLUMNS list in
synapse/_scripts/synapse_port_db.py. This tells the port script to cast
the integer value from SQLite to a boolean before writing the value to the
postgres database.
event_id's can be considered globally unique although there has been a lot of
debate on this topic in places like
MSC2779 and
MSC2848 which
has no resolution yet (as of 2022-09-01). There are several places in Synapse
and even in the Matrix APIs like GET /_matrix/federation/v1/event/{eventId}
where we assume that event IDs are globally unique.
When scoping event_id in a database schema, it is often nice to accompany it
with room_id (PRIMARY KEY (room_id, event_id) and a FOREIGN KEY(room_id) REFERENCES rooms(room_id)) which makes flexible lookups easy. For example it
makes it very easy to find and clean up everything in a room when it needs to be
purged (no need to use sub-select query or join from the events table).
A note on collisions: In room versions 1 and 2 it's possible to end up with
two events with the same event_id (in the same or different rooms). After room
version 3, that can only happen with a hash collision, which we basically hope
will never happen (SHA256 has a massive big key space).
Some migrations need to be performed gradually. A prime example of this is anything
which would need to do a large table scan — including adding columns, indices or
NOT NULL constraints to non-empty tables — such a migration should be done as a
background update where possible, at least on Postgres.
We can afford to be more relaxed about SQLite databases since they are usually
used on smaller deployments and SQLite does not support the same concurrent
DDL operations as Postgres.
We also typically insist on having at least one Synapse version's worth of
backwards compatibility, so that administrators can roll back Synapse if an upgrade
did not go smoothly.
This sometimes results in having to plan a migration across multiple versions
of Synapse.
This section includes an example and may include more in the future.
This example illustrates how you would introduce a new column, write data into it
based on data from an old column and then drop the old column.
We are aiming for semantic equivalence to:
ALTER TABLE mytable ADD COLUMN new_column INTEGER;
UPDATE mytable SET new_column = old_column * 100;
ALTER TABLE mytable ALTER COLUMN new_column ADD CONSTRAINT NOT NULL;
ALTER TABLE mytable DROP COLUMN old_column;
SCHEMA_VERSION = S + 1
SCHEMA_COMPAT_VERSION = ... # unimportant at this stage
Changes:
ALTER TABLE mytable ADD COLUMN new_column INTEGER;
Invariants:
old_column is read by Synapse and written to by Synapse.
new_column is written to by Synapse.
Notes:
new_column can't have a NOT NULL NOT VALID constraint yet, because the previous Synapse version did not write to the new column (since we haven't bumped the SCHEMA_COMPAT_VERSION yet, we still need to be compatible with the previous version).
SCHEMA_VERSION = S + 2
SCHEMA_COMPAT_VERSION = S + 1 # this signals that we can't roll back to a time before new_column existed
Changes:
On Postgres, add a NOT VALID constraint to ensure new rows are compliant. SQLite does not have such a construct, but it would be unnecessary anyway since there is no way to concurrently perform this migration on SQLite.
ALTER TABLE mytable ADD CONSTRAINT CHECK new_column_not_null (new_column IS NOT NULL) NOT VALID;
Start a background update to perform migration: it should gradually run e.g.
UPDATE mytable SET new_column = old_column * 100 WHERE 0 < mytable_id AND mytable_id <= 5;
This background update is technically pointless on SQLite, but you must schedule it anyway so that the portdb script to migrate to Postgres still works.
Upon completion of the background update, you should run VALIDATE CONSTRAINT on Postgres to turn the NOT VALID constraint into a valid one.
ALTER TABLE mytable VALIDATE CONSTRAINT new_column_not_null;
This will take some time but does NOT hold an exclusive lock over the table.
Invariants:
old_column is read by Synapse and written to by Synapse.
new_column is written to by Synapse and new rows always have a non-NULL value in this field.
Notes:
If you wish, you can convert the CHECK (new_column IS NOT NULL) to a NOT NULL constraint free of charge in Postgres by adding the NOT NULL constraint and then dropping the CHECK constraint, because Postgres can statically verify that the NOT NULL constraint is implied by the CHECK constraint without performing a table scan.
It might be tempting to make version N + 2 redundant by moving the background update to N + 1 and delaying adding the NOT NULL constraint to N + 3, but that would mean the constraint would always be validated in the foreground in N + 3. Whereas if the N + 2 step is kept, the migration in N + 3 would be fast in the happy case.
SCHEMA_VERSION = S + 3
SCHEMA_COMPAT_VERSION = S + 1 # we can't roll back to a time before new_column existed
Changes:
(Postgres) Update the table to populate values of new_column in case the background update had not completed. Additionally, VALIDATE CONSTRAINT to make the check fully valid.
-- you ideally want an index on `new_column` or e.g. `(new_column) WHERE new_column IS NULL` first, or perhaps you can find a way to skip this if the `NOT NULL` constraint has already been validated.
UPDATE mytable SET new_column = old_column * 100 WHERE new_column IS NULL;
-- this is a no-op if it already ran as part of the background update
ALTER TABLE mytable VALIDATE CONSTRAINT new_column_not_null;
(SQLite) Recreate the table by precisely following the 12-step procedure for SQLite table schema changes.
During this table rewrite, you should recreate new_column as NOT NULL and populate any outstanding NULL values at the same time.
Unfortunately, you can't drop old_column yet because it must be present for compatibility with the Postgres schema, as needed by portdb.
(Otherwise you could do this all in one go with SQLite!)
Invariants:
old_column is written to by Synapse (but no longer read by Synapse!).
new_column is read by Synapse and written to by Synapse. Moreover, all rows have a non-NULL value in this field, as guaranteed by a schema constraint.
Notes:
We can't drop old_column yet, or even stop writing to it, because that would break a rollback to the previous version of Synapse.
Application code can now rely on new_column being populated. The remaining steps are only motivated by the wish to clean-up old columns.
SCHEMA_VERSION = S + 4
SCHEMA_COMPAT_VERSION = S + 3 # we can't roll back to a time before new_column was entirely non-NULL
Invariants:
old_column exists but is not written to or read from by Synapse.
new_column is read by Synapse and written to by Synapse. Moreover, all rows have a non-NULL value in this field, as guaranteed by a schema constraint.
Notes:
We can't drop old_column yet because that would break a rollback to the previous version of Synapse. TODO: It may be possible to relax this and drop the column straight away as long as the previous version of Synapse detected a rollback occurred and stopped attempting to write to the column. This could possibly be done by checking whether the database's schema compatibility version was S + 3.
It can be desirable to implement "experimental" features which are disabled by
default and must be explicitly enabled via the Synapse configuration. This is
applicable for features which:
Are unstable in the Matrix spec (e.g. those defined by an MSC that has not yet been merged).
Developers are not confident in their use by general Synapse administrators/users
(e.g. a feature is incomplete, buggy, performs poorly, or needs further testing).
Note that this only really applies to features which are expected to be desirable
to a broad audience. The module infrastructure should
instead be investigated for non-standard features.
Guarding experimental features behind configuration flags should help with some
of the following scenarios:
Ensure that clients do not assume that unstable features exist (failing
gracefully if they do not).
Unstable features do not become de-facto standards and can be removed
aggressively (since only those who have opted-in will be affected).
Ease finding the implementation of unstable features in Synapse (for future
removal or stabilization).
Ease testing a feature (or removal of feature) due to enabling/disabling without
code changes. It also becomes possible to ask for wider testing, if desired.
Experimental configuration flags should be disabled by default (requiring Synapse
administrators to explicitly opt-in), although there are situations where it makes
sense (from a product point-of-view) to enable features by default. This is
expected and not an issue.
It is not a requirement for experimental features to be behind a configuration flag,
but one should be used if unsure.
New experimental configuration flags should be added under the experimental
configuration key (see the synapse.config.experimental file) and either explain
(briefly) what is being enabled, or include the MSC number.
Synapse uses a variety of third-party Python packages to function as a homeserver.
Some of these are direct dependencies, listed in pyproject.toml under the
[tool.poetry.dependencies] section. The rest are transitive dependencies (the
things that our direct dependencies themselves depend on, and so on recursively.)
We maintain a locked list of all our dependencies (transitive included) so that
we can track exactly which version of each dependency appears in a given release.
See here
for discussion of why we wanted this for Synapse. We chose to use
poetry to manage this locked list; see
this comment
for the reasoning.
The locked dependencies get included in our "self-contained" releases: namely,
our docker images and our debian packages. We also use the locked dependencies
in development and our continuous integration.
Separately, our "broad" dependencies—the version ranges specified in
pyproject.toml—are included as metadata in our "sdists" and "wheels" uploaded
to PyPI. Installing from PyPI or from
the Synapse source tree directly will not use the locked dependencies; instead,
they'll pull in the latest version of each package available at install time.
[[package]]
name = "phonenumbers"
version = "8.12.44"
description = "Python version of Google's common library for parsing, formatting, storing and validating international phone numbers."
category = "main"
optional = false
python-versions = "*"
The lockfile also includes a
cryptographic checksum
of the sdists and wheels provided for this version of phonenumbers.
direnv is a tool for activating environments in your
shell inside a given directory. Its support for poetry is unofficial (a
community wiki recipe only), but works solidly in our experience. We thoroughly
recommend it for daily use. To use it:
Install direnv - it's likely
packaged for your system already.
Teach direnv about poetry. The shell config here
needs to be added to ~/.config/direnv/direnvrc (or more generally $XDG_CONFIG_HOME/direnv/direnvrc).
Mark the synapse checkout as a poetry project: echo layout poetry > .envrc.
Convince yourself that you trust this .envrc configuration and project.
Then formally confirm this to direnv by running direnv allow.
Then whenever you navigate to the synapse checkout, you should be able to run
e.g. mypy instead of poetry run mypy; python instead of
poetry run python; and your shell commands will automatically run in the
context of poetry's venv, without having to run poetry shell beforehand.
# Stop the current virtualenv if active
$ deactivate
# Remove all of the files from the current environment.
# Don't worry, even though it says "all", this will only
# remove the Poetry virtualenvs for the current project.
$ poetry env remove --all
# Reactivate Poetry shell to create the virtualenv again
$ poetry shell
# Install everything again
$ poetry install --extras all
Use poetry run cmd args when you need the python virtualenv context.
To avoid typing poetry run all the time, you can run poetry shell
to start a new shell in the poetry virtualenv context. Within poetry shell,
python, pip, mypy, trial, etc. are all run inside the project virtualenv
and isolated from the rest o the system.
Roughly speaking, the translation from a traditional virtualenv is:
env/bin/activate -> poetry shell, and
deactivate -> close the terminal (Ctrl-D, exit, etc.)
See also the direnv recommendation above, which makes poetry run and
poetry shell unnecessary.
# Current env only
poetry env info
# All envs: this allows you to have e.g. a poetry managed venv for Python 3.7,
# and another for Python 3.10.
poetry env list --full-path
poetry run pip list
Note that poetry show describes the abstract lock file rather than your
on-disk environment. With that said, poetry show --tree can sometimes be
useful.
ought to do the trick. Alternatively, manually update pyproject.toml and
poetry lock --no-update. Include the updated pyproject.toml and poetry.lock
files in your commit.
to use the latest version of packagename in the locked environment, without
affecting the broad dependencies listed in the wheel.
There doesn't seem to be a way to do this whilst locking a specific version of
packagename. We can workaround this (crudely) as follows:
poetry add packagename==1.2.3
# This should update pyproject.lock.
# Now undo the changes to pyproject.toml. For example
# git restore pyproject.toml
# Get poetry to recompute the content-hash of pyproject.toml without changing
# the locked package versions.
poetry lock --no-update
Either way, include the updated poetry.lock file in your commit.
Synapse uses Dependabot to keep the poetry.lock and Cargo.lock file
up-to-date with the latest releases of our dependencies. The changelog check is
omitted for Dependabot PRs; the release script will include them in the
changelog.
When reviewing a dependabot PR, ensure that:
the lockfile changes look reasonable;
the upstream changelog file (linked in the description) doesn't include any
breaking changes;
continuous integration passes.
In particular, any updates to the type hints (usually packages which start with types-)
should be safe to merge if linting passes.
The minimum version of poetry supported by Synapse is 1.3.2.
It can also be useful to check the version of poetry-core in use. If you've
installed poetry with pipx, try pipx runpip poetry list | grep poetry-core.
Poetry caches a bunch of information about packages that isn't readily available
from PyPI. (This is what makes poetry seem slow when doing the first
poetry install.) Try poetry cache list and poetry cache clear --all <name of cache> to see if that fixes things.
Sometimes, requests take a long time to service and clients disconnect
before Synapse produces a response. To avoid wasting resources, Synapse
can cancel request processing for select endpoints marked with the
@cancellable decorator.
Synapse makes use of Twisted's Deferred.cancel() feature to make
cancellation work. The @cancellable decorator does nothing by itself
and merely acts as a flag, signalling to developers and other code alike
that a method can be cancelled.
Check that the endpoint method, and any async functions in its call
tree handle cancellation correctly. See
Handling cancellation correctly
for a list of things to look out for.
Add the @cancellable decorator to the on_GET/POST/PUT/DELETE
method. It's not recommended to make non-GET methods cancellable,
since cancellation midway through some database updates is less
likely to be handled correctly.
There are two stages to cancellation: downward propagation of a
cancel() call, followed by upwards propagation of a CancelledError
out of a blocked await.
Both Twisted and asyncio have a cancellation mechanism.
When Synapse starts handling a request, it runs the async method
responsible for handling it using defer.ensureDeferred, which returns
a Deferred. For example:
When a client disconnects early, Synapse checks for the presence of the
@cancellable decorator on on_GET. Since on_GET is cancellable,
Deferred.cancel() is called on the Deferred from
defer.ensureDeferred, ie. request. Twisted knows which Deferredrequest is waiting on and passes the cancel() call on to d.
The Deferred being waited on, d, may have its own handling for
cancel() and pass the call on to other Deferreds.
Eventually, a Deferred handles the cancel() call by resolving itself
with a CancelledError.
In general, when writing code that might be subject to cancellation, two
things must be considered:
The effect of CancelledErrors raised out of awaits.
The effect of Deferreds being cancel()ed.
Examples of code that handles cancellation incorrectly include:
try-except blocks which swallow CancelledErrors.
Code that shares the same Deferred, which may be cancelled, between
multiple requests.
Code that starts some processing that's exempt from cancellation, but
uses a logging context from cancellable code. The logging context
will be finished upon cancellation, while the uncancelled processing
is still using it.
Some common patterns are listed below in more detail.
Most functions in Synapse are relatively straightforward from a
cancellation standpoint: they don't do anything with Deferreds and
purely call and await other async functions.
An async function handles cancellation correctly if its own code
handles cancellation correctly and all the async function it calls
handle cancellation correctly. For example:
do_two_things handles cancellation correctly if do_something and
do_something_else handle cancellation correctly.
That is, when checking whether a function handles cancellation
correctly, its implementation and all its async function calls need to
be checked, recursively.
As check_something is not async, it does not need to be checked.
Because Twisted's CancelledErrors are Exceptions, it's easy to
accidentally catch and suppress them. Care must be taken to ensure that
CancelledErrors are allowed to propagate upwards.
broadcasts cancel() calls to every Deferred being waited on.
wraps the first exception it sees in a FirstError.
Together, this means that CancelledErrors will be wrapped in
a FirstError unless unwrapped. Such FirstErrors are liable to be
swallowed, so they must be unwrapped.
If a function creates a Deferred, the effect of cancelling it must be considered. Deferreds that get shared are likely to have unintended behaviour when cancelled.
Bad:
cache: Dict[str, Deferred[None]] = {}
def wait_for_room(room_id: str) -> Deferred[None]:
deferred = cache.get(room_id)
if deferred is None:
deferred = Deferred()
cache[room_id] = deferred
# `deferred` can have multiple waiters.
# All of them will observe a `CancelledError`
# if any one of them is cancelled.
return make_deferred_yieldable(deferred)
# Request 1
await wait_for_room("!aAAaaAaaaAAAaAaAA:matrix.org")
# Request 2
await wait_for_room("!aAAaaAaaaAAAaAaAA:matrix.org")
Good:
cache: Dict[str, Deferred[None]] = {}
def wait_for_room(room_id: str) -> Deferred[None]:
deferred = cache.get(room_id)
if deferred is None:
deferred = Deferred()
cache[room_id] = deferred
# `deferred` will never be cancelled now.
# A `CancelledError` will still come out of
# the `await`.
# `delay_cancellation` may also be used.
return make_deferred_yieldable(stop_cancellation(deferred))
# Request 1
await wait_for_room("!aAAaaAaaaAAAaAaAA:matrix.org")
# Request 2
await wait_for_room("!aAAaaAaaaAAAaAaAA:matrix.org")
Good:
cache: Dict[str, List[Deferred[None]]] = {}
def wait_for_room(room_id: str) -> Deferred[None]:
if room_id not in cache:
cache[room_id] = []
# Each request gets its own `Deferred` to wait on.
deferred = Deferred()
cache[room_id]].append(deferred)
return make_deferred_yieldable(deferred)
# Request 1
await wait_for_room("!aAAaaAaaaAAAaAaAA:matrix.org")
# Request 2
await wait_for_room("!aAAaaAaaaAAAaAaAA:matrix.org")
Some async functions may kick off some async processing which is
intentionally protected from cancellation, by stop_cancellation or
other means. If the async processing inherits the logcontext of the
request which initiated it, care must be taken to ensure that the
logcontext is not finished before the async processing completes.
Bad:
cache: Optional[ObservableDeferred[None]] = None
async def do_something_else(
to_resolve: Deferred[None]
) -> None:
await ...
logger.info("done!")
to_resolve.callback(None)
async def do_something() -> None:
if not cache:
to_resolve = Deferred()
cache = ObservableDeferred(to_resolve)
# `do_something_else` will never be cancelled and
# can outlive the `request-1` logging context.
run_in_background(do_something_else, to_resolve)
await make_deferred_yieldable(cache.observe())
with LoggingContext("request-1"):
await do_something()
Good:
cache: Optional[ObservableDeferred[None]] = None
async def do_something_else(
to_resolve: Deferred[None]
) -> None:
await ...
logger.info("done!")
to_resolve.callback(None)
async def do_something() -> None:
if not cache:
to_resolve = Deferred()
cache = ObservableDeferred(to_resolve)
run_in_background(do_something_else, to_resolve)
# We'll wait until `do_something_else` is
# done before raising a `CancelledError`.
await make_deferred_yieldable(
delay_cancellation(cache.observe())
)
else:
await make_deferred_yieldable(cache.observe())
with LoggingContext("request-1"):
await do_something()
OK:
cache: Optional[ObservableDeferred[None]] = None
async def do_something_else(
to_resolve: Deferred[None]
) -> None:
await ...
logger.info("done!")
to_resolve.callback(None)
async def do_something() -> None:
if not cache:
to_resolve = Deferred()
cache = ObservableDeferred(to_resolve)
# `do_something_else` will get its own independent
# logging context. `request-1` will not count any
# metrics from `do_something_else`.
run_as_background_process(
"do_something_else",
do_something_else,
to_resolve,
)
await make_deferred_yieldable(cache.observe())
with LoggingContext("request-1"):
await do_something()
To help track the processing of individual requests, synapse uses a
'log context' to track which request it is handling at any given
moment. This is done via a thread-local variable; a logging.Filter is
then used to fish the information back out of the thread-local variable
and add it to each log record.
Logcontexts are also used for CPU and database accounting, so that we
can track which requests were responsible for high CPU use or database
activity.
The synapse.logging.context module provides facilities for managing
the current log context (as well as providing the LoggingContextFilter
class).
Asynchronous functions make the whole thing complicated, so this document describes
how it all works, and how to write code which follows the rules.
In this document, "awaitable" refers to any object which can be awaited. In the context of
Synapse, that normally means either a coroutine or a Twisted
Deferred.
In the absence of any asynchronous voodoo, things are simple enough. As with
any code of this nature, the rule is that our function should leave
things as it found them:
from synapse.logging import context # omitted from future snippets
def handle_request(request_id):
request_context = context.LoggingContext()
calling_context = context.set_current_context(request_context)
try:
request_context.request = request_id
do_request_handling()
logger.debug("finished")
finally:
context.set_current_context(calling_context)
def do_request_handling():
logger.debug("phew") # this will be logged against request_id
LoggingContext implements the context management methods, so the above
can be written much more succinctly as:
def handle_request(request_id):
with context.LoggingContext() as request_context:
request_context.request = request_id
do_request_handling()
logger.debug("finished")
def do_request_handling():
logger.debug("phew")
Awaitables break the linear flow of code so that there is no longer a single entry point
where we should set the logcontext and a single exit point where we should remove it.
Consider the example above, where do_request_handling needs to do some
blocking operation, and returns an awaitable:
async def handle_request(request_id):
with context.LoggingContext() as request_context:
request_context.request = request_id
await do_request_handling()
logger.debug("finished")
In the above flow:
The logcontext is set
do_request_handling is called, and returns an awaitable
handle_request awaits the awaitable
Execution of handle_request is suspended
So we have stopped processing the request (and will probably go on to
start processing the next), without clearing the logcontext.
To circumvent this problem, synapse code assumes that, wherever you have
an awaitable, you will want to await it. To that end, wherever
functions return awaitables, we adopt the following conventions:
Rules for functions returning awaitables:
If the awaitable is already complete, the function returns with the
same logcontext it started with.
If the awaitable is incomplete, the function clears the logcontext
before returning; when the awaitable completes, it restores the
logcontext before running any callbacks.
That sounds complicated, but actually it means a lot of code (including
the example above) "just works". There are two cases:
If do_request_handling returns a completed awaitable, then the
logcontext will still be in place. In this case, execution will
continue immediately after the await; the "finished" line will
be logged against the right context, and the with block restores
the original context before we return to the caller.
If the returned awaitable is incomplete, do_request_handling clears
the logcontext before returning. The logcontext is therefore clear
when handle_requestawaits the awaitable.
Once do_request_handling's awaitable completes, it will reinstate
the logcontext, before running the second half of handle_request,
so again the "finished" line will be logged against the right context,
and the with block restores the original context.
As an aside, it's worth noting that handle_request follows our rules
though that only matters if the caller has its own logcontext which it
cares about.
The following sections describe pitfalls and helpful patterns when
implementing these rules.
Whenever you get an awaitable back from a function, you should await on
it as soon as possible. Do not pass go; do not do any logging; do not
call any other functions.
async def fun():
logger.debug("starting")
await do_some_stuff() # just like this
coro = more_stuff()
result = await coro # also fine, of course
return result
Provided this pattern is followed all the way back up to the callchain
to where the logcontext was set, this will make things work out ok:
provided do_some_stuff and more_stuff follow the rules above, then
so will fun.
It's all too easy to forget to await: for instance if we forgot that
do_some_stuff returned an awaitable, we might plough on regardless. This
leads to a mess; it will probably work itself out eventually, but not
before a load of stuff has been logged against the wrong context.
(Normally, other things will break, more obviously, if you forget to
await, so this tends not to be a major problem in practice.)
Of course sometimes you need to do something a bit fancier with your
awaitable - not all code follows the linear A-then-B-then-C pattern.
Notes on implementing more complex patterns are in later sections.
Most of the time, an awaitable comes from another synapse function.
Sometimes, though, we need to make up a new awaitable, or we get an awaitable
back from external code. We need to make it follow our rules.
The easy way to do it is by using context.make_deferred_yieldable. Suppose we want to implement
sleep, which returns a deferred which will run its callbacks after a
given number of seconds. That might look like:
# not a logcontext-rules-compliant function
def get_sleep_deferred(seconds):
d = defer.Deferred()
reactor.callLater(seconds, d.callback, None)
return d
That doesn't follow the rules, but we can fix it by calling it through
context.make_deferred_yieldable:
Sometimes you want to fire off a chain of execution, but not wait for
its result. That might look a bit like this:
async def do_request_handling():
await foreground_operation()
# *don't* do this
background_operation()
logger.debug("Request handling complete")
async def background_operation():
await first_background_step()
logger.debug("Completed first step")
await second_background_step()
logger.debug("Completed second step")
The above code does a couple of steps in the background after
do_request_handling has finished. The log lines are still logged
against the request_context logcontext, which may or may not be
desirable. There are two big problems with the above, however. The first
problem is that, if background_operation returns an incomplete
awaitable, it will expect its caller to await immediately, so will have
cleared the logcontext. In this example, that means that 'Request
handling complete' will be logged without any context.
The second problem, which is potentially even worse, is that when the
awaitable returned by background_operation completes, it will restore
the original logcontext. There is nothing waiting on that awaitable, so
the logcontext will leak into the reactor and possibly get attached to
some arbitrary future operation.
There are two potential solutions to this.
One option is to surround the call to background_operation with a
PreserveLoggingContext call. That will reset the logcontext before
starting background_operation (so the context restored when the
deferred completes will be the empty logcontext), and will restore the
current logcontext before continuing the foreground process:
async def do_request_handling():
await foreground_operation()
# start background_operation off in the empty logcontext, to
# avoid leaking the current context into the reactor.
with PreserveLoggingContext():
background_operation()
# this will now be logged against the request context
logger.debug("Request handling complete")
Obviously that option means that the operations done in
background_operation would be not be logged against a logcontext
(though that might be fixed by setting a different logcontext via a
with LoggingContext(...) in background_operation).
The second option is to use context.run_in_background, which wraps a
function so that it doesn't reset the logcontext even when it returns
an incomplete awaitable, and adds a callback to the returned awaitable to
reset the logcontext. In other words, it turns a function that follows
the Synapse rules about logcontexts and awaitables into one which behaves
more like an external function --- the opposite operation to that
described in the previous section. It can be used like this:
async def do_request_handling():
await foreground_operation()
context.run_in_background(background_operation)
# this will now be logged against the request context
logger.debug("Request handling complete")
This is really a variation of the fire-and-forget problem above, in that
we are firing off a1 and a2 without awaiting on them. The difference
is that we now have third-party code attached to their callbacks. Anyway
either technique given in the Fire-and-forget
section will work.
So, option one: reset the logcontext before starting the operations to
be gathered:
async def do_request_handling():
with PreserveLoggingContext():
a1 = operation1()
a2 = operation2()
result = await defer.gatherResults([a1, a2])
In this case particularly, though, option two, of using
context.run_in_background almost certainly makes more sense, so that
operation1 and operation2 are both logged against the original
logcontext. This looks like:
It turns out that our logcontext rules do not play nicely with awaitable
chains which get orphaned and garbage-collected.
Imagine we have some code that looks like this:
listener_queue = []
def on_something_interesting():
for d in listener_queue:
d.callback("foo")
async def await_something_interesting():
new_awaitable = defer.Deferred()
listener_queue.append(new_awaitable)
with PreserveLoggingContext():
await new_awaitable
Obviously, the idea here is that we have a bunch of things which are
waiting for an event. (It's just an example of the problem here, but a
relatively common one.)
Now let's imagine two further things happen. First of all, whatever was
waiting for the interesting thing goes away. (Perhaps the request times
out, or something even more interesting happens.)
Secondly, let's suppose that we decide that the interesting thing is
never going to happen, and we reset the listener queue:
So, both ends of the awaitable chain have now dropped their references,
and the awaitable chain is now orphaned, and will be garbage-collected at
some point. Note that await_something_interesting is a coroutine,
which Python implements as a generator function. When Python
garbage-collects generator functions, it gives them a chance to
clean up by making the await (or yield) raise a GeneratorExit
exception. In our case, that means that the __exit__ handler of
PreserveLoggingContext will carefully restore the request context, but
there is now nothing waiting for its return, so the request context is
never cleared.
To reiterate, this problem only arises when both ends of a awaitable
chain are dropped. Dropping the the reference to an awaitable you're
supposed to be awaiting is bad practice, so this doesn't
actually happen too much. Unfortunately, when it does happen, it will
lead to leaked logcontexts which are incredibly hard to track down.
We'd like to be able to split some of the work that synapse does into
multiple python processes. In theory multiple synapse processes could
share a single postgresql database and we'd scale up by running more
synapse processes. However much of synapse assumes that only one process
is interacting with the database, both for assigning unique identifiers
when inserting into tables, notifying components about new updates, and
for invalidating its caches.
So running multiple copies of the current code isn't an option. One way
to run multiple processes would be to have a single writer process and
multiple reader processes connected to the same database. In order to do
this we'd need a way for the reader process to invalidate its in-memory
caches when an update happens on the writer. One way to do this is for
the writer to present an append-only log of updates which the readers
can consume to invalidate their caches and to push updates to listening
clients or pushers.
Synapse already stores much of its data as an append-only log so that it
can correctly respond to /sync requests so the amount of code changes
needed to expose the append-only log to the readers should be fairly
minimal.
Information about how the tcp replication module is structured, including how
the classes interact, can be found in
synapse/replication/tcp/__init__.py
Synapse has a concept of "streams", which are roughly described in id_generators.py.
Generally speaking, streams are a series of notifications that something in Synapse's database has changed that the application might need to respond to.
For example:
The events stream reports new events (PDUs) that Synapse creates, or that Synapse accepts from another homeserver.
The account data stream reports changes to users' account data.
The to-device stream reports when a device has a new to-device message.
It is very helpful to understand the streams mechanism when working on any part of Synapse that needs to respond to changes—especially if those changes are made by different workers.
To that end, let's describe streams formally, paraphrasing from the docstring of AbstractStreamIdGenerator.
A stream is an append-only log T1, T2, ..., Tn, ... of facts1 which grows over time.
Only "writers" can add facts to a stream, and there may be multiple writers.
Each fact has an ID, called its "stream ID".
Readers should only process facts in ascending stream ID order.
Roughly speaking, each stream is backed by a database table.
It should have a stream_id (or similar) bigint column holding stream IDs, plus additional columns as necessary to describe the fact.
Typically, a fact is expressed with a single row in its backing table.2
Within a stream, no two facts may have the same stream_id.
Aside. Some additional notes on streams' backing tables.
The backing tables may have other uses.
> For example, the events table serves backs the events stream, and is read when processing new events.
> But old rows are read from the table all the time, whenever Synapse needs to lookup some facts about an event.
Rich suspects that sometimes the stream is backed by multiple tables, so the stream proper is the union of those tables.
Stream writers can "reserve" a stream ID, and then later mark it as having being completed.
Stream writers need to track the completion of each stream fact.
In the happy case, completion means a fact has been written to the stream table.
But unhappy cases (e.g. transaction rollback due to an error) also count as completion.
Once completed, the rows written with that stream ID are fixed, and no new rows
will be inserted with that ID.
For any given stream reader (including writers themselves), we may define a per-writer current stream ID:
A current stream ID for a writer W is the largest stream ID such that
all transactions added by W with equal or smaller ID have completed.
Similarly, there is a "linear" notion of current stream ID:
A "linear" current stream ID is the largest stream ID such that
all facts (added by any writer) with equal or smaller ID have completed.
Because different stream readers A and B learn about new facts at different times, A and B may disagree about current stream IDs.
Put differently: we should think of stream readers as being independent of each other, proceeding through a stream of facts at different rates.
The above definition does not give a unique current stream ID, in fact there can
be a range of current stream IDs. Synapse uses both the minimum and maximum IDs
for different purposes. Most often the maximum is used, as its generally
beneficial for workers to advance their IDs as soon as possible. However, the
minimum is used in situations where e.g. another worker is going to wait until
the stream advances past a position.
NB. For both senses of "current", that if a writer opens a transaction that never completes, the current stream ID will never advance beyond that writer's last written stream ID.
For single-writer streams, the per-writer current ID and the linear current ID are the same.
Both senses of current ID are monotonic, but they may "skip" or jump over IDs because facts complete out of order.
Example.
Consider a single-writer stream which is initially at ID 1.
Treat it as a collection of distinct single-writer streams, one
for each writer.
Treat it as a single stream.
The single stream (option 2) is conceptually simpler, and easier to represent (a single stream id).
However, it requires each reader to know about the entire set of writers, to ensures that readers don't erroneously advance their current stream position too early and miss a fact from an unknown writer.
In contrast, multiple parallel streams (option 1) are more complex, requiring more state to represent (map from writer to stream id).
The payoff for doing so is that readers can "peek" ahead to facts that completed on one writer no matter the state of the others, reducing latency.
Note that a multi-writer stream can be viewed in both ways.
For example, the events stream is treated as multiple single-writer streams (option 1) by the sync handler, so that events are sent to clients as soon as possible.
But the background process that works through events treats them as a single linear stream.
Another useful example is the cache invalidation stream.
The facts this stream holds are instructions to "you should now invalidate these cache entries".
We only ever treat this as a multiple single-writer streams as there is no important ordering between cache invalidations.
(Invalidations are self-contained facts; and the invalidations commute/are idempotent).
track their current position (i.e. its own per-writer stream ID).
their facts currently awaiting completion.
At startup,
the current position of that writer can be found by querying the database (which suggests that facts need to be written to the database atomically, in a transaction); and
there are no facts awaiting completion.
To reserve a stream ID, call nextval on the appropriate postgres sequence.
To write a fact to the stream: insert the appropriate rows to the appropriate backing table.
To complete a fact, first remove it from your map of facts currently awaiting completion.
Then, if no earlier fact is awaiting completion, the writer can advance its current position in that stream.
Upon doing so it should emit an RDATA message3, once for every fact between the old and the new stream ID.
Readers need to track the current position of every writer.
At startup, they can find this by contacting each writer with a REPLICATE message,
requesting that all writers reply describing their current position in their streams.
Writers reply with a POSITION message.
To learn about new facts, readers should listen for RDATA messages and process them to respond to the new fact.
The RDATA itself is not a self-contained representation of the fact;
readers will have to query the stream tables for the full details.
Readers must also advance their record of the writer's current position for that stream.
In a nutshell: we have an append-only log with a "buffer/scratchpad" at the end where we have to wait for the sequence to be linear and contiguous.
1
we use the word fact here for two reasons.
Firstly, the word "event" is already heavily overloaded (PDUs, EDUs, account data, ...) and we don't need to make that worse.
Secondly, "fact" emphasises that the things we append to a stream cannot change after the fact.
2
A fact might be expressed with 0 rows, e.g. if we opened a transaction to persist an event, but failed and rolled the transaction back before marking the fact as completed.
In principle a fact might be expressed with 2 or more rows; if so, each of those rows should share the fact's stream ID.
3
This communication used to happen directly with the writers over TCP;
nowadays it's done via Redis's Pubsub.
Previously the workers used an HTTP long poll mechanism to get updates
from the master, which had the problem of causing a lot of duplicate
work on the server. This TCP protocol replaces those APIs with the aim
of increased efficiency.
The protocol is based on fire and forget, line based commands. An
example flow would be (where '>' indicates master to worker and
'<' worker to master flows):
The example shows the server accepting a new connection and sending its identity
with the SERVER command, followed by the client server to respond with the
position of all streams. The server then periodically sends RDATA commands
which have the format RDATA <stream_name> <instance_name> <token> <row>, where
the format of <row> is defined by the individual streams. The
<instance_name> is the name of the Synapse process that generated the data
(usually "master"). We expect an RDATA for every row in the DB.
Error reporting happens by either the client or server sending an ERROR
command, and usually the connection will be closed.
Since the protocol is a simple line based, its possible to manually
connect to the server using a tool like netcat. A few things should be
noted when manually using the protocol:
The federation stream is only available if federation sending has
been disabled on the main process.
The server will only time connections out that have sent a PING
command. If a ping is sent then the connection will be closed if no
further commands are received within 15s. Both the client and
server protocol implementations will send an initial PING on
connection and ensure at least one command every 5s is sent (not
necessarily PING).
RDATA commands usually include a numeric token, however if the
stream has multiple rows to replicate per token the server will send
multiple RDATA commands, with all but the last having a token of
batch. See the documentation on commands.RdataCommand for
further details.
The basic structure of the protocol is line based, where the initial
word of each line specifies the command. The rest of the line is parsed
based on the command. For example, the RDATA command is defined as:
Both sides are expected to send at least one command every 5s or so, and
should send a PING command if necessary. If either side do not receive
a command within e.g. 15s then the connection should be closed.
Because the server may be connected to manually using e.g. netcat, the
timeouts aren't enabled until an initial PING command is seen. Both
the client and server implementations below send a PING command
immediately on connection to ensure the timeouts are enabled.
This ensures that both sides can quickly realize if the tcp connection
has gone and handle the situation appropriately.
If the server sends messages faster than the client can consume them the
server will first buffer a (fairly large) number of commands and then
disconnect the client. This ensures that we don't queue up an unbounded
number of commands in memory and gives us a potential opportunity to
squawk loudly. When/if the client recovers it can reconnect to the
server and ask for missed messages.
In general the replication stream should be considered an unreliable
transport since e.g. commands are not resent if the connection
disappears.
The exception to that are the replication streams, i.e. RDATA commands,
since these include tokens which can be used to restart the stream on
connection errors.
The client should keep track of the token in the last RDATA command
received for each stream so that on reconnection it can start streaming
from the correct place. Note: not all RDATA have valid tokens due to
batching. See RdataCommand for more details.
On receipt of a POSITION command clients should check if they have missed any
updates, and if so then fetch them out of band. Sent in response to a
REPLICATE command (but can happen at any time).
The POSITION command includes the source of the stream. Currently all streams
are written by a single process (usually "master"). If fetching missing
updates via HTTP API, rather than via the DB, then processes should make the
request to the appropriate process.
Two positions are included, the "new" position and the last position sent respectively.
This allows servers to tell instances that the positions have advanced but no
data has been written, without clients needlessly checking to see if they
have missed any updates. Instances will only fetch stuff if there is a gap between
their current position and the given last position.
The cache invalidation stream is used to inform workers when they need
to invalidate any of their caches in the data store. This is done by
streaming all cache invalidations done on master down to the workers,
assuming that any caches on the workers also exist on the master.
Each individual cache invalidation results in a row being sent down
replication, which includes the cache name (the name of the function)
and they key to invalidate. For example:
However, there are times when a number of caches need to be invalidated
at the same time with the same key. To reduce traffic we batch those
invalidations into a single poke by defining a special cache name that
workers understand to mean to expand to invalidate the correct caches.
Currently the special cache names are declared in
synapse/storage/_base.py and are:
cs_cache_fake ─ invalidates caches that depend on the current
state
The key idea is described by MSC3706. This allows servers to
request a lightweight response to the federation /send_join endpoint.
This is called a faster join, also known as a partial join. In these
notes we'll usually use the word "partial" as it matches the database schema.
a list of the servers in the room (according to the state before J),
a subset of the state of the room before J,
the full auth chain of that state subset.
Synapse marks the room as partially joined by adding a row to the database table
partial_state_rooms. It also marks the join event J as "partially stated",
meaning that we have neither received nor computed the full state before/after
J. This is done by adding a row to partial_state_events.
DB schema
matrix=> \d partial_state_events
Table "matrix.partial_state_events"
Column │ Type │ Collation │ Nullable │ Default
══════════╪══════╪═══════════╪══════════╪═════════
room_id │ text │ │ not null │
event_id │ text │ │ not null │
matrix=> \d partial_state_rooms
Table "matrix.partial_state_rooms"
Column │ Type │ Collation │ Nullable │ Default
════════════════════════╪════════╪═══════════╪══════════╪═════════
room_id │ text │ │ not null │
device_lists_stream_id │ bigint │ │ not null │ 0
join_event_id │ text │ │ │
joined_via │ text │ │ │
matrix=> \d partial_state_rooms_servers
Table "matrix.partial_state_rooms_servers"
Column │ Type │ Collation │ Nullable │ Default
═════════════╪══════╪═══════════╪══════════╪═════════
room_id │ text │ │ not null │
server_name │ text │ │ not null │
Indices, foreign-keys and check constraints are omitted for brevity.
While partially joined to a room, Synapse receives events E from remote
homeservers as normal, and can create events at the request of its local users.
However, we run into trouble when we enforce the checks on an event.
Is a valid event, otherwise it is dropped. For an event to be valid, it
must contain a room_id, and it must comply with the event format of that
room version.
Passes signature checks, otherwise it is dropped.
Passes hash checks, otherwise it is redacted before being processed further.
Passes authorization rules based on the event’s auth events, otherwise it
is rejected.
Passes authorization rules based on the state before the event, otherwise
it is rejected.
Passes authorization rules based on the current state of the room,
otherwise it is “soft failed”.
We can enforce checks 1--4 without any problems.
But we cannot enforce checks 5 or 6 with complete certainty, since Synapse does
not know the full state before E, nor that of the room.
Instead, we make a best-effort approximation.
While the room is considered partially joined, Synapse tracks the "partial
state" before events.
This works in a similar way as regular state:
The partial state before J is that given to us by the partial join response.
The partial state before an event E is the resolution of the partial states
after each of E's prev_events.
If E is rejected or a message event, the partial state after E is the
partial state before E.
Otherwise, the partial state after E is the partial state before E, plus
E itself.
More concisely, partial state propagates just like full state; the only
difference is that we "seed" it with an incomplete initial state.
Synapse records that we have only calculated partial state for this event with
a row in partial_state_events.
While the room remains partially stated, check 5 on incoming events to that
room becomes:
Passes authorization rules based on the resolution between the partial
state before E and E's auth events. If the event fails to pass
authorization rules, it is rejected.
Additionally, check 6 is deleted: no soft-failures are enforced.
While partially joined, the current partial state of the room is defined as the
resolution across the partial states after all forward extremities in the room.
Remark. Events with partial state are not considered
outliers.
Using partial state means the auth checks can fail in a few different ways1.
1
Is this exhaustive?
We may erroneously accept an incoming event in check 5 based on partial state
when it would have been rejected based on full state, or vice versa.
This means that an event could erroneously be added to the current partial
state of the room when it would not be present in the full state of the room,
or vice versa.
Additionally, we may have skipped soft-failing an event that would have been
soft-failed based on full state.
(Note that the discrepancies described in the last two bullets are user-visible.)
This means that we have to be very careful when we want to lookup pieces of room
state in a partially-joined room. Our approximation of the state may be
incorrect or missing. But we can make some educated guesses. If
our partial state is likely to be correct, or
the consequences of our partial state being incorrect are minor,
then we proceed as normal, and let the resync process fix up any mistakes (see
below).
When is our partial state likely to be correct?
It's more accurate the closer we are to the partial join event. (So we should
ideally complete the resync as soon as possible.)
Non-member events: we will have received them as part of the partial join
response, if they were part of the room state at that point. We may
incorrectly accept or reject updates to that state (at first because we lack
remote membership information; later because of compounding errors), so these
can become incorrect over time.
Local members' memberships: we are the only ones who can create join and
knock events for our users. We can't be completely confident in the
correctness of bans, invites and kicks from other homeservers, but the resync
process should correct any mistakes.
Remote members' memberships: we did not receive these in the /send_join
response, so we have essentially no idea if these are correct or not.
In short, we deem it acceptable to trust the partial state for non-membership
and local membership events. For remote membership events, we wait for the
resync to complete, at which point we have the full state of the room and can
proceed as normal.
The partial-state approximation is only a temporary affair. In the background,
synapse beings a "resync" process. This is a continuous loop, starting at the
partial join event and proceeding downwards through the event graph. For each
E seen in the room since partial join, Synapse will fetch
the event ids in the state of the room before E, via
/state_ids;
the event ids in the full auth chain of E, included in the /state_ids
response; and
any events from the previous two bullets that Synapse hasn't persisted, via
`/state.
This means Synapse has (or can compute) the full state before E, which allows
Synapse to properly authorise or reject E. At this point ,the event
is considered to have "full state" rather than "partial state". We record this
by removing E from the partial_state_events table.
[TODO: Does Synapse persist a new state group for the full state
before E, or do we alter the (partial-)state group in-place? Are state groups
ever marked as partially-stated? ]
This scheme means it is possible for us to have accepted and sent an event to
clients, only to reject it during the resync. From a client's perspective, the
effect is similar to a retroactive
state change due to state resolution---i.e. a "state reset".2
2
Clients should refresh caches to detect such a change. Rumour has it that
sliding sync will fix this.
When all events since the join J have been fully-stated, the room resync
process is complete. We record this by removing the room from
partial_state_rooms.
For the time being, the resync process happens on the master worker.
A new replication stream un_partial_stated_room is added. Whenever a resync
completes and a partial-state room becomes fully stated, a new message is sent
into that stream containing the room ID.
When sending out messages during a partial join, we assume our partial state is
accurate and proceed as normal. For this to have any hope of succeeding at all,
our partial state must contain an entry for each of the (type, state key) pairs
specified by the auth rules:
m.room.create
m.room.join_rules
m.room.power_levels
m.room.third_party_invite
m.room.member
The first four of these should be present in the state before J that is given
to us in the partial join response; only membership events are omitted. In order
for us to consider the user joined, we must have their membership event. That
means the only possible omission is the target's membership in an invite, kick
or ban.
The worst possibility is that we locally invite someone who is banned according to
the full state, because we lack their ban in our current partial state. The rest
of the federation---at least, those who are fully joined---should correctly
enforce the membership transition constraints. So any the erroneous invite should be ignored by fully-joined
homeservers and resolved by the resync for partially-joined homeservers.
In more generality, there are two problems we're worrying about here:
We might create an event that is valid under our partial state, only to later
find out that is actually invalid according to the full state.
Or: we might refuse to create an event that is invalid under our partial
state, even though it would be perfectly valid under the full state.
However we expect such problems to be unlikely in practise, because
We trust that the room has sensible power levels, e.g. that bad actors with
high power levels are demoted before their ban.
We trust that the resident server provides us up-to-date power levels, join
rules, etc.
State changes in rooms are relatively infrequent, and the resync period is
relatively quick.
Normally: send out in a fed txn to all HSes in the room.
We only know that some HSes were in the room at some point. Wat do.
Send it out to the list of servers from the first join.
TODO what do we do here if we have full state?
If the prev event was created by us, we can risk sending it to the wrong HS. (Motivation: privacy concern of the content. Not such a big deal for a public room or an encrypted room. But non-encrypted invite-only...)
But don't want to send out sensitive data in other HS's events in this way.
Suppose we discover after resync that we shouldn't have sent out one our events (not a prev_event) to a target HS. Not much we can do.
What about if we didn't send them an event but shouldn't've?
E.g. what if someone joined from a new HS shortly after you did? We wouldn't talk to them.
Could imagine sending out the "Missed" events after the resync but... painful to work out what they should have seen if they joined/left.
Instead, just send them the latest event (if they're still in the room after resync) and let them backfill.(?)
Don't do this currently.
If anyone who has received our messages sends a message to a HS we missed, they can backfill our messages
Gap: rooms which are infrequently used and take a long time to resync.
TODO: needs prose fleshing out. Liase with Matthieu. Explain why /send_join
(Rich was surprised we didn't just create it locally. Answer: to try and avoid
a join which then gets rejected after resync.)
We don't know for sure that any join we create would be accepted.
E.g. the joined user might have been banned; the join rules might have changed in a way that we didn't realise... some way in which the partial state was mistaken.
Instead, do another partial make-join/send-join handshake to confirm that the join works.
Probably going to get a bunch of duplicate state events and auth events.... but the point of partial joins is that these should be small. Many are already persisted = good.
What if the second send_join response includes a different list of reisdent HSes? Could ignore it.
Could even have a special flag that says "just make me a join", i.e. don't bother giving me state or servers in room. Deffo want the auth chain tho.
SQ: wrt device lists it's a lot safer to ignore it!!!!!
What if the state at the second join is inconsistent with what we have? Ignore it?
If type is m.room.member: [...]
>
> 5. If membership is leave:
>
> 1. If the sender matches state_key, allow if and only if that user’s current membership state is invite, join, or knock.
2. [...]
I think this means that (well-formed!) self-leaves are governed entirely by
4.5.1. This means that if we correctly calculate state which says that U is
invited, joined or knocked and include it in the leave's auth events, our event
is accepted by checks 4 and 5 on incoming events.
Passes authorization rules based on the event’s auth events, otherwise
> it is rejected.
Passes authorization rules based on the state before the event, otherwise
> it is rejected.
The only way to fail check 6 is if the receiving server's current state of the
room says that U is banned, has left, or has no membership event. But this is
fine: the receiving server already thinks that U isn't in the room.
Passes authorization rules based on the current state of the room,
> otherwise it is “soft failed”.
For the second point (publishing the leave event), the best thing we can do is
to is publish to all HSes we know to be currently in the room. If they miss that
event, they might send us traffic in the room that we don't care about. This is
a problem with leaving after a "full" join; we don't seek to fix this with
partial joins.
(With that said: there's nothing machine-readable in the /send response. I don't
think we can deduce "destination has left the room" from a failure to /send an
event into that room?)
We can create leave events and can choose what gets included in our auth events,
so we can be sure that we pass check 4 on incoming events. For check 5, we might
have an incorrect view of the state before an event.
The only way we might erroneously think a leave is valid is if
the partial state before the leave has U joined, invited or knocked, but
the full state before the leave has U banned, left or not present,
in which case the leave doesn't make anything worse: other HSes already consider
us as not in the room, and will continue to do so after seeing the leave.
The remaining obstacle is then: can we safely broadcast the leave event? We may
miss servers or incorrectly think that a server is in the room. Or the
destination server may be offline and miss the transaction containing our leave
event.This should self-heal when they see an event whose prev_events descends
from our leave.
Another option we considered was to use federation /send_leave to ask a
fully-joined server to send out the event on our behalf. But that introduces
complexity without much benefit. Besides, as Rich put it,
sending out leaves is pretty best-effort currently
This section covers implementation documentation for various parts of Synapse.
If a developer is planning to make a change to a feature of Synapse, it can be useful for
general documentation of how that feature is implemented to be available. This saves the
developer time in place of needing to understand how the feature works by reading the
code.
Documentation that would be more useful for the perspective of a system administrator,
rather than a developer who's intending to change to code, should instead be placed
under the Usage section of the documentation.
https://fujifish.github.io/samling/samling.html (https://github.com/fujifish/samling) is a great resource for being able to tinker with the
SAML options within Synapse without needing to deploy and configure a complicated software stack.
To make Synapse (and therefore Element) use it:
Use the samling.html URL above or deploy your own and visit the IdP Metadata tab.
Copy the XML to your clipboard.
On your Synapse server, create a new file samling.xml next to your homeserver.yaml with
the XML from step 2 as the contents.
Edit your homeserver.yaml to include:
saml2_config:
sp_config:
allow_unknown_attributes: true # Works around a bug with AVA Hashes: https://github.com/IdentityPython/pysaml2/issues/388
metadata:
local: ["samling.xml"]
Ensure that your homeserver.yaml has a setting for public_baseurl:
public_baseurl: http://localhost:8080/
Run apt-get install xmlsec1 and pip install --upgrade --force 'pysaml2>=4.5.0' to ensure
the dependencies are installed and ready to go.
Restart Synapse.
Then in Element:
Visit the login page and point Element towards your homeserver using the public_baseurl above.
Click the Single Sign-On button.
On the samling page, enter a Name Identifier and add a SAML Attribute for uid=your_localpart.
The response must also be signed.
Click "Next".
Click "Post Response" (change nothing).
You should be logged in.
If you try and repeat this process, you may be automatically logged in using the information you
gave previously. To fix this, open your developer console (F12 or Ctrl+Shift+I) while on the
samling page and clear the site data. In Chrome, this will be a button on the Application tab.
The word "edge" comes from graph theory lingo. An edge is just a connection
between two events. In Synapse, we connect events by specifying their
prev_events. A subsequent event points back at a previous event.
Events are normally sorted by (topological_ordering, stream_ordering) where
topological_ordering is just depth. In other words, we first sort by depth
and then tie-break based on stream_ordering. depth is incremented as new
messages are added to the DAG. Normally, stream_ordering is an auto
incrementing integer, but backfilled events start with stream_ordering=-1 and decrement.
/sync returns things in the order they arrive at the server (stream_ordering).
/messages (and /backfill in the federation API) return them in the order determined by the event graph (topological_ordering, stream_ordering).
The general idea is that, if you're following a room in real-time (i.e.
/sync), you probably want to see the messages as they arrive at your server,
rather than skipping any that arrived late; whereas if you're looking at a
historical section of timeline (i.e. /messages), you want to see the best
representation of the state of the room as others were seeing it at the time.
We mark an event as an outlier when we haven't figured out the state for the
room at that point in the DAG yet. They are "floating" events that we haven't
yet correlated to the DAG.
Outliers typically arise when we fetch the auth chain or state for a given
event. When that happens, we just grab the events in the state/auth chain,
without calculating the state at those events, or backfilling their
prev_events. Since we don't have the state at any events fetched in that
way, we mark them as outliers.
So, typically, we won't have the prev_events of an outlier in the database,
(though it's entirely possible that we might have them for some other
reason). Other things that make outliers different from regular events:
We don't have state for them, so there should be no entry in
event_to_state_groups for an outlier. (In practice this isn't always
the case, though I'm not sure why: see https://github.com/matrix-org/synapse/issues/12201).
We don't record entries for them in the event_edges,
event_forward_extremeties or event_backward_extremities tables.
Since outliers are not tied into the DAG, they do not normally form part of the
timeline sent down to clients via /sync or /messages; however there is an
exception:
A special case of outlier events are some membership events for federated rooms
that we aren't full members of. For example:
invites received over federation, before we join the room
rejections for said invites
knock events for rooms that we would like to join but have not yet joined.
In all the above cases, we don't have the state for the room, which is why they
are treated as outliers. They are a bit special though, in that they are
proactively sent to clients via /sync.
Most-recent-in-time events in the DAG which are not referenced by any other
events' prev_events yet. (In this definition, outliers, rejected events, and
soft-failed events don't count.)
The forward extremities of a room (or at least, a subset of them, if there are
more than ten) are used as the prev_events when the next event is sent.
The "current state" of a room (ie: the state which would be used if we
generated a new event) is, therefore, the resolution of the room states
at each of the forward extremities.
The current marker of where we have backfilled up to and will generally be the
prev_events of the oldest-in-time events we have in the DAG. This gives a starting point when
backfilling history.
Note that, unlike forward extremities, we typically don't have any backward
extremity events themselves in the database - or, if we do, they will be "outliers" (see
above). Either way, we don't expect to have the room state at a backward extremity.
When we persist a non-outlier event, if it was previously a backward extremity,
we clear it as a backward extremity and set all of its prev_events as the new
backward extremities if they aren't already persisted as non-outliers. This
therefore keeps the backward extremities up-to-date.
For every non-outlier event we need to know the state at that event. Instead of
storing the full state for each event in the DB (i.e. a event_id -> state
mapping), which is very space inefficient when state doesn't change, we
instead assign each different set of state a "state group" and then have
mappings of event_id -> state_group and state_group -> state.
A state set is a set of state events; e.g. the input of a state resolution
algorithm is a collection of state sets.
The auth chain of a set of events are all the events' auth events and their
auth events, recursively (i.e. the events reachable by walking the graph induced
by an event's auth events links).
The auth chain difference of a collection of state sets is the union minus the
intersection of the sets of auth chains corresponding to the state sets, i.e an
event is in the auth chain difference if it is reachable by walking the auth
event graph from at least one of the state sets but not from all of the state
sets.
A way of calculating the auth chain difference without calculating the full auth
chains for each state set is to do a parallel breadth first walk (ordered by
depth) of each state set's auth chain. By tracking which events are reachable
from each state set we can finish early if every pending event is reachable from
every state set.
This can work well for state sets that have a small auth chain difference, but
can be very inefficient for larger differences. However, this algorithm is still
used if we don't have a chain cover index for the room (e.g. because we're in
the process of indexing it).
Synapse computes auth chain differences by pre-computing a "chain cover" index
for the auth chain in a room, allowing us to efficiently make reachability queries
like "is event A in the auth chain of event B?". We could do this with an index
that tracks all pairs (A, B) such that A is in the auth chain of B. However, this
would be prohibitively large, scaling poorly as the room accumulates more state
events.
Instead, we break down the graph into chains. A chain is a subset of a DAG
with the following property: for any pair of events E and F in the chain,
the chain contains a path E -> F or a path F -> E. This forces a chain to be
linear (without forks), e.g. E -> F -> G -> ... -> H. Each event in the chain
is given a sequence number local to that chain. The oldest event E in the
chain has sequence number 1. If E has a child F in the chain, then F has
sequence number 2. If E has a grandchild G in the chain, then G has
sequence number 3; and so on.
Synapse ensures that each persisted event belongs to exactly one chain, and
tracks how the chains are connected to one another. This allows us to
efficiently answer reachability queries. Doing so uses less storage than
tracking reachability on an event-by-event basis, particularly when we have
fewer and longer chains. See
In practical terms, the chain cover assigns every event a
chain ID and sequence number (e.g. (5,3)), and maintains a map of links
between events in chains (e.g. (5,3) -> (2,4)) such that A is reachable by B
(i.e. A is in the auth chain of B) if and only if either:
A and B have the same chain ID and A's sequence number is less than B's
sequence number; or
there is a link L between B's chain ID and A's chain ID such that
L.start_seq_no <= B.seq_no and A.seq_no <= L.end_seq_no.
There are actually two potential implementations, one where we store links from
each chain to every other reachable chain (the transitive closure of the links
graph), and one where we remove redundant links (the transitive reduction of the
links graph) e.g. if we have chains C3 -> C2 -> C1 then the link C3 -> C1
would not be stored. Synapse uses the former implementation so that it doesn't
need to recurse to test reachability between chains. This trades-off extra storage
in order to save CPU cycles and DB queries.
An example auth graph would look like the following, where chains have been
formed based on type/state_key and are denoted by colour and are labelled with
(chain ID, sequence number). Links are denoted by the arrows (links in grey
are those that would be remove in the second implementation described above).
Note that we don't include all links between events and their auth events, as
most of those links would be redundant. For example, all events point to the
create event, but each chain only needs the one link from it's base to the
create event.
This index can be used to calculate the auth chain difference of the state sets
by looking at the chain ID and sequence numbers reachable from each state set:
For every state set lookup the chain ID/sequence numbers of each state event
Use the index to find all chains and the maximum sequence number reachable
from each state set.
The auth chain difference is then all events in each chain that have sequence
numbers between the maximum sequence number reachable from any state set and
the minimum reachable by all state sets (if any).
Note that steps 2 is effectively calculating the auth chain for each state set
(in terms of chain IDs and sequence numbers), and step 3 is calculating the
difference between the union and intersection of the auth chains.
Synapse generates 24 character media IDs for content uploaded by local users.
These media IDs consist of upper and lowercase letters and are case-sensitive.
Other homeserver implementations may generate media IDs differently.
Local media is recorded in the local_media_repository table, which includes
metadata such as MIME types, upload times and file sizes.
Note that this table is shared by the URL cache, which has a different media ID
scheme.
When media from a remote homeserver is requested from Synapse, it is assigned
a local filesystem_id, with the same format as locally-generated media IDs,
as described above.
A record of remote media is stored in the remote_media_cache table, which
can be used to map remote MXC URIs (server names and media IDs) to local
filesystem_ids.
When generating previews for URLs, Synapse may download and cache various
resources, including images. These resources are assigned temporary media IDs
of the form yyyy-mm-dd_aaaaaaaaaaaaaaaa, where yyyy-mm-dd is the current
date and aaaaaaaaaaaaaaaa is a random sequence of 16 case-sensitive letters.
The metadata for these cached resources is stored in the
local_media_repository and local_media_repository_url_cache tables.
Resources for URL previews are deleted after a few days.
Synapse maintains room and user statistics in various tables. These can be used
for administrative purposes but are also used when generating the public room
directory.
Synapse has a number of platform dependencies, including Python, Rust,
PostgreSQL and SQLite. This document outlines the policy towards which versions
we support, and when we drop support for versions in the future.
Synapse follows the upstream support life cycles for Python and PostgreSQL,
i.e. when a version reaches End of Life Synapse will withdraw support for that
version in future releases.
A Rust compiler is required to build Synapse from source. For any given release
the minimum required version may be bumped up to a recent Rust version, and so
people building from source should ensure they can fetch recent versions of Rust
(e.g. by using rustup).
It is important for system admins to have a clear understanding of the platform
requirements of Synapse and its deprecation policies so that they can
effectively plan upgrading their infrastructure ahead of time. This is
especially important in contexts where upgrading the infrastructure requires
auditing and approval from a security team, or where otherwise upgrading is a
long process.
By following the upstream support life cycles Synapse can ensure that its
dependencies continue to get security patches, while not requiring system admins
to constantly update their platform dependencies to the latest versions.
For Rust, the situation is a bit different given that a) the Rust foundation
does not generally support older Rust versions, and b) the library ecosystem
generally bump their minimum support Rust versions frequently. In general, the
Synapse team will try to avoid updating the dependency on Rust to the absolute
latest version, but introducing a formal policy is hard given the constraints of
the ecosystem.
On a similar note, SQLite does not generally have a concept of "supported
release"; bugfixes are published for the latest minor release only. We chose to
track Debian's oldstable as this is relatively conservative, predictably updated
and is consistent with the .deb packages released by Matrix.org.
I've been running my homeserver on a cubietruck at home now for some time and am often replying to statements like "you need loads of ram to join large rooms" with "it works fine for me". I thought it might be useful to curate a summary of the issues you're likely to run into to help as a scaling-down guide, maybe highlight these for development work or end up as documentation. It seems that once you get up to about 4x1.5GHz arm64 4GiB these issues are no longer a problem.
This is the main reason people have a poor matrix experience on resource constrained homeservers. Element web will frequently be saying the server is offline while the python process will be pegged at 100% cpu. This feature is used to tell when other users are active (have a client app in the foreground) and therefore more likely to respond, but requires a lot of network activity to maintain even when nobody is talking in a room.
While synapse does have some performance issues with presence #3971, the fundamental problem is that this is an easy feature to implement for a centralised service at nearly no overhead, but federation makes it combinatorial #8055. There is also a client-side config option which disables the UI and idle tracking enable_presence_by_hs_url to blacklist the largest instances but I didn't notice much difference, so I recommend disabling the feature entirely at the server level as well.
Joining a "large", federated room will initially fail with the below message in Element web, but waiting a while (10-60mins) and trying again will succeed without any issue. What counts as "large" is not message history, user count, connections to homeservers or even a simple count of the state events, it is instead how long the state resolution algorithm takes. However, each of those numbers are reasonable proxies, so we can use them as estimates since user count is one of the few things you see before joining.
This is #1211 and will also hopefully be mitigated by peeking matrix-org/matrix-doc#2753 so at least you don't need to wait for a join to complete before finding out if it's the kind of room you want. Note that you should first disable presence, otherwise it'll just make the situation worse #3120. There is a lot of database interaction too, so make sure you've migrated your data from the default sqlite to postgresql. Personally, I recommend patience - once the initial join is complete there's rarely any issues with actually interacting with the room, but if you like you can just block "large" rooms entirely.
Anything that requires modifying the device list #7721 will take a while to propagate, again taking the client "Offline" until it's complete. This includes signing in and out, editing the public name and verifying e2ee. The main mitigation I recommend is to keep long-running sessions open e.g. by using Firefox SSB "Use this site in App mode" or Chromium PWA "Install Element".
Put the below in a new file at /etc/matrix-synapse/conf.d/sbc.yaml to override the defaults in homeserver.yaml.
# Disable presence tracking, which is currently fairly resource intensive
# More info: https://github.com/matrix-org/synapse/issues/9478
use_presence: false
# Set a small complexity limit, preventing users from joining large rooms
# which may be resource-intensive to remain a part of.
#
# Note that this will not prevent users from joining smaller rooms that
# eventually become complex.
limit_remote_rooms:
enabled: true
complexity: 3.0
# Database configuration
database:
# Use postgres for the best performance
name: psycopg2
args:
user: matrix-synapse
# Generate a long, secure password using a password manager
password: hunter2
database: matrix-synapse
host: localhost
Currently the complexity is measured by current_state_events / 500. You can find join times and your most complex rooms like this:
.
