Sick monitoring server

[Vyacheslav]

From time to time we see such alerts from monitoring servers:

CPU 3 is averaging 607.85 interrupts/sec.            No

CPU 1 is averaging 678.09 interrupts/sec.            No

CPU 2 is averaging 681.77 interrupts/sec.            No

CPU 0 is averaging 619.06 interrupts/sec.           

Also, sometimes false treshholds are triggered on AES/AES servers - either snmp is not available or the server itself is not available.

I assume the causes of false alarms are somehow related to interrupts in the CPU. There are also suggestions that somehow do not feel the monitoring server.
I would be grateful for advice on how to eliminate the difficulties.

[SCOTT_BALDWIN]

Hello Vyacheslav,

This issue will likely require Support to perform a Health Check on the server to determine the underlying cause of the interrupts and missing data. I see a case has already been opened, and the Support Engineer will work with you directly.

One of the biggest causes of performance issues on VMs is the Socket and CPU Core settings. Prognosis servers should be configured with at least 4 CPUs per Socket. Using 1 Socket per CPU (VMware default) causes performance issues with Prognosis.

Thank you,

Scott Baldwin