Requires python3 and pip (on Debian/Ubuntu: apt install python3-pip).
Run pip3 install -r requirements.txt to install python dependencies (might be just pip on other distributions)
See ./monitoring.nagios.py --help
The script runs a specified health check on Ceph, if no health check is provided a check on the overall Ceph health is executed.
The simplest way to run this is like this:
./monitoring.nagios.py --host localhost --user admin --pass admin
This will map Ceph health warnings and errors to WARNING and CRITICAL service states in nagios. You can optionally pass a specific Ceph health check as parameter to monitor on a finer granularity. For example, it might be useful to mute the general Ceph health warning during ongoing recovery but still get alerted if an OSD fails.
Ceph offers a lot of health checks, we've found the following to be useful with Nagios/Icinga:
MON_DOWN
MON_CLOCK_SKEW
OSD_DOWN
OSD_NEARFULL
OSD_FULL
OSDMAP_FLAGS
POOL_FULL
POOL_NEAR_FULL
PG_AVAILABILITY
PG_DEGRADED
PG_DAMAGED
OBJECT_UNFOUND
SLOW_OP
MDS_SLOW_METADATA_IO
FS_DEGRADED
MDS_INSUFFICIENT_STANDBY