Nagios Support Forum

How do I increase performance graph frequency from 5 min averages to 1 min? I have already changed the 'Check interval' from 5 to 1.

Thanks,

The short answer is is that there is no benefit to setting it to be that short. The performance graphs are set to sync with MRTG updates, and MRTG updates are set for 5 minutes. The reasoning behind this is MRTG updates and RRD graphs are very I/O intensive, if these were set to be 1 minute, the CPU and I/O usage would be insane. We are looking into way to make bandwidth graphs update in 1 minutes intervals, but for now, 5 minutes is what is suppported.

thanks for the quick reply. You are probably right, most likely I won't be doing this in production.

In my current test case, I am writing my own plugin and running it via check_nrpe. I see performance graphs being generated automatically. Is there a way to generate these graph for 1 minute instead of 5?

mrtg is primarily for switches and routers, as for most other checks, if you set the check_interval at 1mn, it should give you graph data at 1mn intervals. Just be aware of doing this in a large scale, as it will multiply your check volume x5.

I've tried changing the 'Check interval' to 1 but I did not immediately see any changes in the graph.
1. am I suppose to see any changes?
2. where are these data stored? how do I keep track of their size?

Thanks,

am I suppose to see any changes?

You may not see a noticeable difference in the graph. rrdtool determines each time the graph is generated what the graph "resolution" will be, and that somewhat depends on the timeframe of the graph.

where are these data stored? how do I keep track of their size?

rrd files are located at /usr/local/nagios/share/perfdata, and once they are created their filesizes remain static, so they will never grow beyond what their original size is.

Not tested, but in theory the workaround for this problem could be to:

1. Remove current rrd and xml files gathering the performance data (for the services that You want to change the interval to 1 minute).
2. Change the rra settings (in /usr/local/nagios/etc/pnp/rra.cfg) - I'm not sure, but it may require restart of the npcd daemon to take effect.
3. Wait for new performance data to arrive and for PNP/Nagios to create the new rrd/xml files (ensure there are no other services waiting for the first performance data - since this would create new rrd files with this settings too).
4. Change rra config back to original (restart npcd ?) - from this point all new rrd file should be created with standard settings.

After this You should have rrd files that can absorb data in 1 minute intervals - so graphs should reflect this

Take care

Michalux

Yeah I hadn't thought about the rra settings, that might be a workable solution. Thanks for the input!