Re: Average CPU load on Controller VM is high (above average load threshold value 10)
I believe that this check has been introduced in one of the latest NCC (Or been reworked with new metrics), and are usually a False/Positive. If I heard right this check will be disabled in one of the upcoming NCC Updates to address the False/Positive Issues, and then be reworked and reintroduced with proper metrics.
As always, if you are concern of the cluster stability open up a support ticket and have them look at the logs to verify if it’s a False/Positive or not. You can also use “sar –q” on the CVM that the alert is coming from to get a glimpse on the CPU Load.
Depending on the amount of vCPUs that the CVM has, the average value should not be higher than 1.5 x vCPU. So with a standard setup most CVMs come with a vCPU of 8, meaning that the average value should not be higher than 12. A few peaks can be normal due to background process (Curator) and/or higher I/O Load.
I hope that this could shed some light on your question.