add more metrics - Githubissues

gesellix / couchdb-prometheus-exporter

CouchDB stats exporter for Prometheus

MIT License

58 stars 20 forks source link

add more metrics #2

Open gesellix opened 8 years ago

gesellix commented 8 years ago

We currently only provide metrics from the _stats endpoint, but there's more: see http://gws.github.io/munin-plugin-couchdb/guide-to-couchdb-monitoring.html for a nice overview and https://github.com/gws/munin-plugin-couchdb for a reference implementation.
As of CouchDB 2.0 there's a dedicated /_up endpoint available: http://docs.couchdb.org/en/2.0.0/whatsnew/2.0.html
Another source of inspiration are the examples for Netdata: https://github.com/firehol/netdata/pull/2818 (graphs: https://imgur.com/a/4Gz7q, setup: https://gist.github.com/anonymous/e490d40ef208d8ef89396f04eed44fc2)

gesellix commented 7 years ago

Progress so far:

[x] /_stats
[x] /_active_tasks
[x] /<db>
[x] Add availability metric per node (#12)
[x] replication state/progress (#17)
[x] View "staleness" (#32)
[x] /_scheduler/jobs (#34)
[ ] /_scheduler/docs
[ ] /_system (#37)

gesellix commented 7 years ago

Hey @uschtwill would you be interested in testing the new metrics? I added _active_tasks and detailed database disk stats (by adding --databases=db-1,db-2 parameter). Your feedback would be appreciated! :-)

uschtwill commented 7 years ago

Hey @gesellix, I'd love to. But I have to warn you, I am currently up to my nose with work, so it might be a while before I get around to trying it.

gesellix commented 7 years ago

@uschtwill Sounds good! Feedback at any time is better than no feedback ;-)

nevans commented 6 years ago

See also _system: https://couchdb.readthedocs.io/en/latest/api/server/common.html#node-node-name-system. The most obviously important of those metrics are already in _stats, but the message_queues counts can occasionally be interesting for guessing at bottlenecks.