2015-04-23 17:07:49 +02:00
|
|
|
How to monitor Synapse metrics using Prometheus
|
|
|
|
|
===============================================
|
|
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
1. Install prometheus:
|
2015-04-23 17:07:49 +02:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
Follow instructions at http://prometheus.io/docs/introduction/install/
|
2015-04-23 17:07:49 +02:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
2. Enable synapse metrics:
|
2015-04-23 17:07:49 +02:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
Simply setting a (local) port number will enable it. Pick a port.
|
|
|
|
|
prometheus itself defaults to 9090, so starting just above that for
|
|
|
|
|
locally monitored services seems reasonable. E.g. 9092:
|
2015-04-23 17:07:49 +02:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
Add to homeserver.yaml::
|
|
|
|
|
|
|
|
|
|
metrics_port: 9092
|
|
|
|
|
|
|
|
|
|
Also ensure that ``enable_metrics`` is set to ``True``.
|
2017-02-19 23:55:48 +01:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
Restart synapse.
|
2015-04-23 17:07:49 +02:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
3. Add a prometheus target for synapse.
|
|
|
|
|
|
2017-04-21 17:03:32 +02:00
|
|
|
It needs to set the ``metrics_path`` to a non-default value (under ``scrape_configs``)::
|
2016-10-28 14:58:27 +02:00
|
|
|
|
|
|
|
|
- job_name: "synapse"
|
|
|
|
|
metrics_path: "/_synapse/metrics"
|
|
|
|
|
static_configs:
|
2017-04-21 17:03:32 +02:00
|
|
|
- targets: ["my.server.here:9092"]
|
2016-10-31 16:06:52 +01:00
|
|
|
|
2017-02-20 00:06:08 +01:00
|
|
|
If your prometheus is older than 1.5.2, you will need to replace
|
|
|
|
|
``static_configs`` in the above with ``target_groups``.
|
2017-02-20 00:06:45 +01:00
|
|
|
|
|
|
|
|
Restart prometheus.
|
2017-02-20 00:06:08 +01:00
|
|
|
|
2016-10-31 16:06:52 +01:00
|
|
|
Standard Metric Names
|
|
|
|
|
---------------------
|
|
|
|
|
|
|
|
|
|
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 miliseconds.
|
|
|
|
|
|
|
|
|
|
================================== =============================
|
|
|
|
|
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.
|
|
|
|
|
|
2016-11-03 18:04:13 +01:00
|
|
|
==================================== =====================
|
2016-10-31 16:06:52 +01:00
|
|
|
New name Old name
|
2016-11-03 18:04:13 +01:00
|
|
|
------------------------------------ ---------------------
|
|
|
|
|
python_twisted_reactor_pending_calls reactor_pending_calls
|
2016-10-31 16:06:52 +01:00
|
|
|
python_twisted_reactor_tick_time reactor_tick_time
|
2016-11-03 18:04:13 +01:00
|
|
|
==================================== =====================
|
