Faults Super Metric

[JCOEN]

I'm trying to create a super metric that combines vCenter faults and host faults. The idea is I have a main operations screen that shows me if a certain site has faults. I can list them out separately, but prefer to combine into one metric. I've tried many combinations of summing both metrics, using max and maxN, but no luck so far. Any ideas?

[mark_j]

Fault are merely classification type of alerts and triggered by change events picked up by the vCenter adapter. If we're going to try and exclusively quantity them, we'd need a metric/attribute on each resource that specifically counts Fault alert types. We have a metric that count alerts by active status and criticality levels, but nothing that specifically counts Faults. So quantifying them is out of the question. However, you 'could' sum/min/max the badge|Fault metric, which is the Fault badge. What this will do is at least give you an indicator that Faults are, actually, present and tell you what the most severe criticality is.. and with SMs you bubble that number up to wherever you like in the relationship hierarchy.

[JCOEN]

I've been working with the badge|Fault metric to try and get the information I want, but so far I have been unsuccessful with combing vCenter and Host faults. I either get one or the other, depending on the object I'm applying the metric to. I've tried applying the super metric to the vCenter resource and Datacenter resource. I've tried a varying number of combinations for the super metric (sum/max/max) but still am not getting a combined view.

[mark_j · Accepted solution]

You just need to get a little more creative with the SM formula.   Here's something to think about:

max([$This:A804,maxN($RK19:A1553,3)])

and enumerated...

max([This Resource: Badge|Fault,maxN(Host System: Badge|Fault,3)])

Yes, you can use these brackets.

Apply this SM to vCenter resources. What this does is get the vCenter, and the highest Fault value of all of the hosts 3 degrees of separation under the vCenter... then the SM will take the value of whichever is highest.. the Fault score for vCenter or the highest Fault of your hosts. Magic!

#supermetric #super metric #conditional logic

[JCOEN]

Thanks Mark, I'll give that a try and let you know. That is definitely more creative than I was getting.

Is there some advanced documentation on super metrics, or is it a trial and error thing?

[mark_j]

Advanced docs on SMs.. not so much. The best formal stuff I've seen out there has been basic. Right now if you want the good stuff, PSO can assist with that. PSO has answers by virtue of the sheer volume of deployments and encountering many wacky use-case scenarios.. so it requires a certain degree of creativity beyond the written word sometimes.

Then there's always this community, which I contribute to for this very purpose. Once things hit this community, people can Google it

[JCOEN]

Figured as much. Thanks for the help. Glad to see you can use brackets. This example will definitely help me create SMs in the future.