Alarm shaping

[Daveyd]

Is there a way to "shape" alarms, kind of like vCenter OPS does, to not trigger when common events happen on a regular basis on a VM. Say I have a VM that runs a certain routine once an hour (24 times a day), every day that causes a 15 minute spike in CPU utilization. I would get 24 error alarms and potentially 24 alarm resets every day. Excluding the VM from this alarm is not an option since I still want to receive CPU utilization events.

Is there a way for Veeam to "learn" the pattern of these type of common events and potentially not alarm on them or something to that effect?

[Vitaliy S.]

Hi Dave, since you have these alarms happening on a regular basis, I believe it should be possible to suppress these alarms on certain hours?

[Daveyd]

This is probably a dumb question, but can you supress specific alarms multiple times a day?

[Vitaliy S.]

It's not a dumb question! Currently you can specify only 1 time interval to suppress alarms. Out of curiosity, what kind of activity is performed once or more times a day that triggers CPU Utilization alarm?

[Daveyd]

It's not a dumb question! Currently you can specify only 1 time interval to suppress alarms. Out of curiosity, what kind of activity is performed once or more times a day that triggers CPU Utilization alarm?

I thought it was a dumb questions since I only saw that you could only surpress alarms once and thought I was missing something

I have a VM that runs multiple reports at once every hour. The compiling of the reports is CPU intensive since the reports are all run at once (they have to be). CPU utilization during that timeframe is a little above 90% and lasts for about 15 minutes thus tripping the alarm. This happens several time a day. Looking at the repeat count, there are 106 instances of the alarm tripping the getting resolved in the past few weeks.

As a workaround, I guess I could exclude that VM from the default CPU utilization alarm and create a new CPU utilization alarm just assigned to that (those) particular VMs and adjust the error thresholds?

[Vitaliy S.]

Yes, as a workaround this would certainly work, thanks for clarification on your use case, that helps.