add more metrics

[gesellix]

• 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: CouchDB python module (from Apache CouchDB maintainers) netdata/netdata#2818 (graphs: https://imgur.com/a/4Gz7q, setup: https://gist.github.com/anonymous/e490d40ef208d8ef89396f04eed44fc2)

[gesellix]

Progress so far:

• /_stats
• /_active_tasks
• /<db>
• Add availability metric per node (Add availability metric per node #12)
• replication state/progress (add replication status metrics #17)
• View "staleness" (Add metric to observe staleness of Views #32)
• /_scheduler/jobs (Observe scheduler jobs #34)
• /_scheduler/docs
• /_system (V2 metrics #37)

[gesellix]

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]

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]

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

[nevans]

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.