Merge pull request #1187 from matrix-org/paul/metrics-howto

Update documentation about exported prometheus metrics
This commit is contained in:
Paul Evans 2016-10-31 18:33:32 +00:00 committed by GitHub
commit 47ed4d84bb
1 changed files with 34 additions and 25 deletions

View File

@ -15,36 +15,45 @@ How to monitor Synapse metrics using Prometheus
Restart synapse Restart synapse
3: Check out synapse-prometheus-config 3: Add a prometheus target for synapse. It needs to set the ``metrics_path``
https://github.com/matrix-org/synapse-prometheus-config to a non-default value::
4: Add ``synapse.html`` and ``synapse.rules`` - job_name: "synapse"
The ``.html`` file needs to appear in prometheus's ``consoles`` directory, metrics_path: "/_synapse/metrics"
and the ``.rules`` file needs to be invoked somewhere in the main config static_configs:
file. A symlink to each from the git checkout into the prometheus directory - targets:
might be easiest to ensure ``git pull`` keeps it updated. "my.server.here:9092"
5: Add a prometheus target for synapse Standard Metric Names
This is easiest if prometheus runs on the same machine as synapse, as it can ---------------------
then just use localhost::
global: { As of synapse version 0.18.2, the format of the process-wide metrics has been
rule_file: "synapse.rules" changed to fit prometheus standard naming conventions. Additionally the units
} have been changed to seconds, from miliseconds.
job: { ================================== =============================
name: "synapse" 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
================================== =============================
target_group: { The python-specific counts of garbage collector performance have been renamed.
target: "http://localhost:9092/"
}
}
6: Start prometheus:: =========================== ======================
New name Old name
--------------------------- ----------------------
python_gc_time reactor_gc_time
python_gc_unreachable_total reactor_gc_unreachable
python_gc_counts reactor_gc_counts
=========================== ======================
./prometheus -config.file=prometheus.conf The twisted-specific reactor metrics have been renamed.
7: Wait a few seconds for it to start and perform the first scrape, ==================================== =================
then visit the console: New name Old name
------------------------------------ -----------------
http://server-where-prometheus-runs:9090/consoles/synapse.html python_twisted_reactor_pending_calls reactor_tick_time
python_twisted_reactor_tick_time reactor_tick_time
==================================== =================