Nagios Support Forum

I'm trying to get my head around the performance data, and I've got a few questions...

1. Should performance data for deleted hosts / services be purged automatically? '/usr/local/nagios/share/perfdata/' still contains data from hosts we've deleted. Should we delete this manually, or is it something that should be done automatically (if so, how frequently?)?

2. I've created the following custom 'rra.cfg' config:
This should record data as follows:

1 Minute - 30 Days
5 Minutes - 90 Days
30 Minutes - 180 Days
6 Hours - 1 Year
1 Day - 3 Years

These RRD files are going to be a fair bit bigger than the 400k default. I was just wondering whether you've any data on how well Nagios will cope with larger RRD files? Is there an upper limit? We have rrdcached configured, so hopefully the I/O should be manageable.

3. When reading the documentation for creating plugins, I've noticed there's an option to specify 'critical'/'threshold' targets in the performance output: 'label'=value[UOM];[warn];[crit];. When I've written plugins previously, I've passed the thresholds to the plugin, and had the plugin calculate whether the thresholds were breached, and pass a return code back to Nagios. Is this not the preferred way of doing it? Does one method override the other?

Sorry for the bombardment of questions, we just want to make sure our Nagios installation is perfect!

Many thanks,

Gavin

Gavin wrote:1. Should performance data for deleted hosts / services be purged automatically? '/usr/local/nagios/share/perfdata/' still contains data from hosts we've deleted. Should we delete this manually, or is it something that should be done automatically (if so, how frequently?)?

Nagios XI doesn't delete RRD because sometimes Items get deleted and then readded, or a host gets renamed and the new files need to be migrated to the new hostname.

Gavin wrote:These RRD files are going to be a fair bit bigger than the 400k default. I was just wondering whether you've any data on how well Nagios will cope with larger RRD files? Is there an upper limit? We have rrdcached configured, so hopefully the I/O should be manageable.

These RRD files are going to be a lot bigger, XI shouldn't have trouble with them, but make sure you have enough disk space to manage this....

Gavin wrote: 3. When reading the documentation for creating plugins, I've noticed there's an option to specify 'critical'/'threshold' targets in the performance output: 'label'=value[UOM];[warn];[crit];. When I've written plugins previously, I've passed the thresholds to the plugin, and had the plugin calculate whether the thresholds were breached, and pass a return code back to Nagios. Is this not the preferred way of doing it? Does one method override the other?

This is correct, but if you include the threshold in the perfdata, Nagios can draw a line on the graph at the critical value...