1,) We have a data center with 11 AC units across multiple rows of racks.
2.) I have one of five custom pollers in an advanced alert that triggers on numeric temperature value when inlet temperature (MaxRackInletUS) goes above a certain value in Fahrenheit, and assigned 'Critical' in Custom poller thresholds, which I have showing in views in dashboard.
3.) This works fine indivdually when one or two AC's exceed thresholds, but we had a catastrophic outage of the Chiller unit that feeds all 11 AC's, so it flooded alerts coming out from all of them. So, I want to set up a poller threshold dependency.
Is it possible to make one of the Chillers OID for say 'offline' or 'shutdown' a parent custom poller value set to 'Critical' threshold, and have the 11 AC units (MaxRackInletUS) custom poller as children of the Chillers custom poller, so if the Chiller is pinging 'UP', but poller value goes 'Critical' due 'offline' etc, then one alert is sent out from Chiller, suppressing all 11 Child AC units triggering, thereby avoiding a flood? Do I have to produce a 'DOWN' status in order for this to work, as I've done with Nodes?
I don't have access to the Chiller yet, so in the interim I simply created a CHILD group of all but one of the AC units of the assigned MaxRackInletUS pollers, then created a GROUP for use as PARENT from the excluded AC unit of the same poller.
My quick fix for now was to set up one AC unit as parent, and set temp inlet threshold lower than the other to trigger earlier before others do, thereby suppressing the others in case the Chiller gets cranky again.
Haven't tested this yet, but I appreciate any input on design logic/capability, as I know the ability to set custom device poller thresholds were just recently added with recent NPM releases.
Here's what the advanced alert looks like, followed by child and parent configs.