Nagios Support Forum

I recently installed Nagios xi on a Dell Poweredge 2950 with these specs:

Dual Quad Core Xeon E5410 Processors
8GB RAM
6 - 145GB 15K SAS Hard Drives in a RAID 10 Array
GB Nic

I am currently up to 566 checks with approximately 500 of those being run at 1 minute intervals. Last night I received my first warning from the localhost cpu load check, It was around 5% - 7%.

I can change the cpu load check numbers so that I don't get warnings, but what do most people see for cpu server load on their Nagios servers? Is this normal behavior when running this many checks at this short of an interval?

Maybe. It all depends on the frequency and overhead of the checks. Simple ICMP checks will take much less load than a database check that needs to login and do a query. Have you noticed an undue increase in the load recently? What type of checks are you running?

localhost cpu load check, It was around 5% - 7%.

CPU load isn't actually measured in %, since you have 8 CPU cores on your machine, normal usage of that many cores will range from 5.0-16.0. You can probably get away with setting your warning threshold at 20.0, and critical at 25.0 or 30.0. This article explains CPU load fairly well.

http://blog.scoutapp.com/articles/2009/ ... d-averages

Understand also that spikes at the 1mn intervals are pretty normal, so you mainly want to worry if the CPU load after is above your thresholds on the 15mn average. That means the machine really is probably overtaxed.

The load increase was expected since I have added a lot more checks. A majority of the checks are for status and bandwidth on network switches, check_rrdtraf and check_ifoperstatus. Thank you for the excellent link to Linux CPU loads article, that was very helpful and answered my questions.

Thank's for letting us know that you found a resolution. Switch and router checks do like to eat up system resources but I guess that is the nature of the beast. Closing thread as resolved.