nagios-ceph-perf

Nagios Plugin and PNP4Nagios Template for Acquiring Ceph Perf Data.

why this and not the ceph distributed one?

It's considerably more detailed information than the one ceph hosts, including read/write throughput information and other detailed info a human could want.

plugin args

usage: check_ceph_perf [-h] [-q] [-i ID] [-k KEYRING] [-e EXEC] [-s SAMPLES]
                       [-d SAMPLE_DELAY] [--generic_level GENERIC_LEVEL]
                       [--mon_generic] [--mon_generic_perf]
                       [--mon_storage_perf] [--mon_store_stats_perf]
                       [--osd_generic] [--osd_perf] [--pg_perf_storage]
                       [--pg_perf_states] [--pg_perf_throughput]
                       [--pg_perf_ops]

'ceph perf' nagios plugin.

optional arguments:
  -h, --help            show this help message and exit
  -q, --quick           quick output configuration (generic + | +
                        pg_perf_storage + pg_perf_througput + pg_perf_ops +
                        osd_perf), one sample unless defined elsewhere.
  -i ID, --id ID        ceph client id
  -k KEYRING, --keyring KEYRING
                        ceph client keyring file
  -e EXEC, --exec EXEC  ceph executable path [/usr/bin/ceph]
  -s SAMPLES, --samples SAMPLES
                        samples to take before providing output
  -d SAMPLE_DELAY, --sample_delay SAMPLE_DELAY
                        time between taking each sample in seconds
  --generic_level GENERIC_LEVEL
                        sets verbosity of output for the generic print (Non-
                        perf). 0 shows status, 1 shows quorum, 2 shows epoch,
                        3 shows fsid, 4 shows round.
  --mon_generic         render mon generic data (host, ip, ranking, updated)
  --mon_generic_perf    render mon perf data (latency, skew)
  --mon_storage_perf    render mon perf data (kb_avail,_total,_used)
  --mon_store_stats_perf
                        render mon perf data (bytes_log,_misc,_sst,_total)
  --osd_generic         render generic osd data (nearfull, full), defaults to
                        0
  --osd_perf            render osd perf data (num_osds, _up_osds,_in_osds
  --pg_perf_storage     render pg perf data (bytes_avail, _total, _used)
  --pg_perf_states      render pg perf states (Renders all possible states
                        [large amount of data])
  --pg_perf_throughput  render pg perf throughput (write/read bytes per
                        second)
  --pg_perf_ops         render pg perf ops (ops)

notes

This will conflict with the ceph_perf nagios module, keep that in mind when testing.