Dear SolarWinds,
I have come across the need of configuring CPU resource alerts per core.
Most of the times, CPU alert can be dependent on Warning/Critical thresholds. Thanks to this approach, we can only edit thresholds and avoid editing multiple alerts, reports, widgets etc.
Since there is no threshold available per core, such alerts need to be set up manually with written percentage in "actual trigger condition" (i.e. higher than 90% for at least 5 min).
First use case:
This would be very useful for processes without multithreading ability, as those processes would consume 100% only in one core (in case of loop).
In servers with multiple cores, this would not trigger the overall threshold value for CPU, so we could potentially not discover such issue (without manual alerts).
I understand that in SAM module, we could use some custom alerts for the processes themselves, however in NPM we do no have such option.
Second use case:
I have noticed that for some network devices, operating with 100% core usage could impact production.
The best example is Palo Alto, where SNMP is showing us two cores. However those are for two different purposes - CPU Data and CPU Management statistics.
(panCPU_DataPlane & panCPU_MgmtPlane OID names)
In such case, achieving 100% only on one of them (i.e. CPU Data), would be a critical situation.
Thanks for your help.
Kind regards,
Marcin Kazmierczak.
