Huge /var/spool/icinga2/perfdata directory

Hello all,

Recently I discovered that our Icinga2 installation was eating a lot of diskspace. After checking with ncdu I discovered that the /var/spool/icinga2/perfdata was almost 78GB.

After reading some forum threads I discovered that the perfdata feature isn’t always required. How to tell if it’s required or not? I’m using icinga2 with icinga2web with influxdb integration and grafana. The documentation doesn’t seem to mention it anywhere specifically to require it, but want ot be sure before turning it off

Disabled features: command compatlog debuglog elasticsearch gelf livestatus opentsdb statusdata syslog
Enabled features: api checker graphite ido-mysql influxdb mainlog notification perfdata

Used web modules: businessprocess, director, grafana, idoreports, incubator, ipl, monitoring, pdfexport, reactbundle, reporting, vspheredb

Is there an alternative (safe) way to clean these logs? Is it OK if you just manually delete these log files?

Afaik the perfdata feature and the files are only needed for pnp4nagios as a graphing solution.
As these files should normally be deleted after they have been processed, it seems there is nothing using the files.
As you have the graphite and influxdb features enabled I would say you can safely disable the perfdata feature and delete the files.
I do so with my graphite setups, and it works well.

edit: here is the corresponding entry from the documentation.
So the files/feature is only needed for pnp4nagios or graphios

1 Like

Thank you for the fast response. I’ve disabled it then and I’ll see what it’ll do :slight_smile:

I’ve rewritten that part in git master for 2.11 lately, since the mistake happens too often scrolling down the docs. perfdata is something for legacy 1.x reasons, and then moved at the very bottom.


1 Like