How does Workload badge handle custom applications?

[MattG]

I want to have an accurate workload score for a custom application.  

Custom objects have an associated Workload score OOB.   How does vCops handle custom applications with Super Metrics?  What does vCops rollup into the custom applications Workload?   Would I need to create  a Custom App Super Metric as a KPI for it to rollup into the Application's Workload score?

Thanks,

-MattG

[mark_j]

An Application Container doesn't have a workload metric by default. Only specific RKs get workloads calculated from certain adapters.. which doesn't include Tiers or Applications.

When you say "Custom objects", what are you specifically referring to?

[MattG]

When I say "Custom Objects"  I mean Application/Tiers.   I just double-checked and my apps do not have a workload associated with them by default.  There is a Health Score. 

How do the Health Score of an Application work?  What does it track?   I have VM with SuperMetric KPIs,  do they automatically affect the Health Score at the Application level?    If the Health Score a reliable method to track custom Application health (rolling up all the children into the Health Score).   If so,  how far down does health look?  Would it include SAN adapter metrics in its score?

Thanks,

-MattG

[mark_j]

Containers such as Application and Tier RKs use anomalies (DTs) and KPIs to calculate health scores.

If you have a Resource under an Application or Tier that has a KPI, that will also create a ripple effect, if you will, or anomaly counts up to parent resources (apps and tiers) when there is a KPI breach.

An Application would include SAN adapter health if.. a App is populated with VMs, that is running on datastores.. and those datastores have relationships created to SAN resources underneath.

It's all about the relationships.

And yes, it is reliable because the methodology scales to just about anything (leveraging anomalies and KPIs).

[MattG]

How are the Anomalies weighted for DT of KPI vs Non-KPI metrics?   For example, Do KPIs affect Health 2x more than non-KPIs?

Also how are objects weighted for Health scores?  If a VM has 3 child objects (Datastore, SAN, Hyperic),  how does vCops weigh the child objects?   1/3 each?

Thanks,

-MattG

[mark_j]

I don't recall what the exactly anomaly translation is to to each KPI, but you'll see the anomaly count spike and raise above the anomaly watermark for certain on a particular resource.

So a non-KPI metric counts 1.. while a KPI counts less-so an absolute value, but I observe to be a % of the total anomalies or % of the watermark for that resource. I don't recall what the exact verbiage is on how the anomaly count is quantified for a KPI.. I'll keep an eye out for it.

When you think about how this roles up hill, the quantity of anomalies of the resource as a percentage of the aggregate is what will make it stand out (and associated KPI / notification alert). A KPI will show very much at a specific resource, then less-so at a parent, then less-so at the parent of that. If you want something more direct or 'pointed' at a parent resource of a child, use a supermetric as a roll-up, aggregate, or min/max as an SLA.