CPU Load versus CPU count

[DigitalEmu]

I am rebuilding the Nagios alert structure for our office running mostly Linux machines.

We currently have every single machine alerting at a certain CPU Load which is inefficient.

Is there a way in the nagios.cfg file to set the cpu load warning to = cpu count?

Basically I want to alarm when the CPU Load goes higher than the CPU count and do it in a way that I can push the cfg file to all my machines with no extra setup.

Thanks in Advance!

[slansing]

No the nagios.cfg does not control that aspect of things, alerting is done by setting thresholds which are determined based on what the plugin you are checking against returns, so this would have to be edited in the plugin itself. Are you installing Nagios on all of your linux systems..? You could just use a lightweight client such as NRPE to execute the nagios plugins remotely and return the data to your central Nagios server, however if that is not possible.. you could use software such as Puppet to push the config files to your other servers.

[DigitalEmu]

pushing the config or software would be done by puppet yes.

You will have to excuse me I am a bit of a Nagios n00b

So is there a plugin that will do this natively or am I going to need to build my own?

[slansing]

Well what I was pointing out is that you could install the NRPE agent on those systems, rather than installing a full fledged Nagios instance, that way your central Nagios server could reap the data from each client, that is the way most tend to set up their monitoring environment, it could save you quite a bit of time and resources on the systems.

I'm not sure if we have any plugins available that will do that natively, though they would be listed on our exchange, here:

http://exchange.nagios.org/index.php?op ... inux%20cpu