As you know, whenever you look at load average on a Linux Machine, you should
take into account the number of cores.
This means that a load average (1 minute) value of "4" on a quad core is good;
instead if you have a single core,it is not.
This is important when you set a threshold for the load average and, on nagios
exchange, there are plugins that take care of this aspect (e.g:
Using standard nrpe, the check_load plugin has the -r option (Divide the load
averages by the number of CPUs)
In my opinion having a plugin like this is very nice because you can set the
same threshold for all linux machines without taking care whenever someone add
or remove cpu especially on virtual machines.
How does check_MK work? Is possible to do something like that? If not how do
you manage cpu_load threshold on linux machines?
checkmk-en mailing list