I'm familiarizing myself with monitoring in vROPS. Among other things, I'm trying to sort out how to alert on VMs that are powered off but ignoring those that are legitimately powered off. We have some dashboards that have been created for us and we've been informed that we can create groups of resources that are excluded from the reporting. I'm curious about alerts and notifications, though. Is there a reasonable way to ignore VMs that might be powered off for good reason when generating alerts and sending notifications? Do the exclusion groups for dashboard reports also affect alerts and notifications?
I think that you can do it by create or modifying policies.
Here you can find a quick info how to use policies:
I will ask you a question that will put into context the problem with this type of alert. This is not just a problem with vrops but all alerting tools out there.
Say you invite me to your office and put me in front of a vcenter how would i know if a VM is powered of legitimately or not just by looking at it?
That is the problem you will have with alerting. How will you tell vrops that a VM has been powered off legitimately and not alert on it. The power state in vrops is either powered on or off.
You could use a script that powers off vms legitimately and add a custom property to a VM that intern adds a VM to a custom group based on the property but group membership is only updated every 20 mins.
You could script the power off and have it put the VM into maintenance mode in vrops. That will suppress all alerts from that VM. Down side is all monitoring of that vm stops and you have to make sure to use a script to power it on and take it out of mm
A easy place to start is to ask what is a non legitimately powered off VM mean. There may be other alerts that may help
How will you tell vrops that a VM has been powered off legitimately and not alert on it.
I think you can use vSphere tags and create custom group for these VMs. And then try to disable alert only for this group.