[check_postgres] nagios perfdata, consistent per-database stats

CSS css at morefoo.com
Mon May 14 04:55:48 UTC 2012


Hello,

I'm having a problem with graphing some of the performance data using pnp4nagios when using check_postgres_locks.

The plugin properly checks the number of locks, regardless of which database the locks exist in.  However, the performance data output only includes databases that currently have a lock.  This makes graphing hard because each time the plugin is run, the number of datasources that are reported might be 1 or it might be 12.  pnp4nagios (and I assume other graphing add-ons for nagios) uses perfdata to create an rrd datastore from which it draws graphs, and it generally assumes that each key in the perfdata output will be consistent on each run.

Would it be possible to add a (global?) switch in check_postgres to always output a full list of databases in perfdata output even if the count of whatever is being monitored per-database is zero?  I see there's an option to limit the total number of perfdata items ("--perflimit=x"), I'm looking for something that's sort of the opposite of that…

Any pointers, or should I just dig around in the code?

Thanks,

Charles


More information about the Check_postgres mailing list