check cpu load, which one to use?

[ekrengel]

So we currently use what I guess is the built in nagios plugin to check the CPU load on linux/redhat servers. But it appears to only check the cpu load average, like how you would see it in the top command. But it does not alert when a certain process is spiking for more than 5 minutes. So with the top command you see the load average in the top right hand corner, but when looking at the processes you see that one is spiking to 100 percent so we don't get alerted for that.

Does anyone know which script on nagios exchange will do this? We may just have if configured wrong too...

[agriffin]

You can use the check_procs plugin to check CPU load on processes (something like `check_procs -m CPU -w 70 -c 85`). You can't specify a timeframe, like "it's been spiking for 5 minutes," but you could set up notifications on this check so that you only get an alert if it hasn't recovered in 5 minutes.

[ekrengel]

For the "CPU" after the -m, is that for a specific process or just for any process that spikes over 70 and 85? Just not sure if you were using that as an example.

[agriffin]

That would check for any process over 70 or 85. You can also filter it to one or a few processes, but I don't think that's you want.