Are there best practice defaults for system thresholds?

[#1]

Hi;

Are there best practice defaults for system thresholds? For example:

. CPU temperature warning threshold

. CPU temperature critical threshold

. CPU load warning threshold

. CPU load critical thresholod

. Memory usage thresholds

. Swap usage thresholds

Kindly

Wasfi

[#2 · Accepted solution]

These values originate from the manual - NIOS_8.6_Documentation.pdf

It is less about warning/critical, more about trap trigger/reset.   For a warning/critical, you would SNMP poll and warning would be above the reset, critical above the trigger ...

CPU temperature warning threshold/CPU temperature critical threshold - There does not appear to be a Trap for this.   ibCPUTemperature is available from SNMP, and there is of course the LOM if enabled (example in doc).   I would look at the Intel ARK document for finer details depending on the model of CPU you have.

CPU load warning threshold/CPU load critical threshold - trigger 81%, reset 70%

Memory usage thresholds - trigger 90%, reset 80%

Swap usage thresholds - trigger 50%, reset 30%