processor load is too high for windows server 2012

[bertaudmarc]

I receive too much alerts for a windows server 2012 installed in a virtual machine: processor load is too high
The standard configuration is system.cpu.load[percpu,avg1].avg(5m) > 5

Some people say in the forum this configuration is not adapted for windows installed in a vm.

I don't know by what I can modify the configuration for receiving less messages but for keeping alert realistic.

[Atsushi]

The fact that the value of the load average is high means that the waiting state of the program is occurring as much.
In other words, the response of the server should be slow.

If there is no problem even if the processing is delayed, please adjust the threshold of the trigger.

The default value of the template is just a reference value,
and you should adjust the threshold according to the environment and usage.
It does not matter whether it is VM or not.

[bertaudmarc]

The question is what is the right value for the threshold !
The question was for windows server 2012 specialists who use Zabbix

[Atsushi]

There is no right value for the threshold applicable to all systems.

If it is okay for the response to be slow, you should raise the threshold.
If processing delay is not allowed, you should change to a higher performance VM.

It depends on the performance required of the system.
Please adjust according to your system and service.

[onallion]

Basically, that depends on the amount of cores.

So for example on a 2-core server, a value of up to 2 would be fine (no more than 1.7 is ideal). Any more than that, and it means processes are waiting.

Now, depending on what the server is doing that might not be a problem you want to be alerted about. For a database I would imagine it would be critical.

If you have an item that collects the amount of cores on a host, you could create a calculated trigger that would adjust itself accordingly

[bertaudmarc]

Ok. I added one core to my vm and will check the load.