Metric Thresholds in vROPs 7

[RobertWLSEG]

Can anyone confirm if you can configure a maximum value that vROPs will store for a calculated metric?

Here is my scenario. I am having to re-write a whole bunch of reports that were written in vROPs 6 which used Super Metrics extensively (these were created to cater for our own thresholds and limits). We embarked on a rightsizing exersize for our estate and one of the metrics we used was cpu|size.recommendation which showed the number of vCPU that an oversized VM could be reduced to and still cope with the same workload. The figure always seemed super agressive, 16 vCPU down to 2 vCPU was not uncommon. With vROPs 7, this metric is no longer available, instead we are presented with a metric which displays as  summary|oversized| Virtual CPUs which gives us the number of CPU that a VM can be reduced by (subtly different to the old metric as that showed what it could be reduced to). Simple enough to transpose in a SM (Provisioned vCPU - summary|oversized| Virtual CPUs = correct number of vCPU for the VM based on its workload).

So I've written the SM and written the report, but have noticed that in no single case is the value of summary|oversized| Virtual CPUs > 50% of the original provisioned vCPU. Now my real question is, has there been a decision taken that caps the value of this field or is it configurable in the Policy to be more or less aggressive?

Any help/guidance/advice gratefully received.

[sxnxr · Accepted solution]

This is by design now. Vrops will only tell you do reduce by a max of 50%. This is like a guide rail in case the more aggressive reduction had an adverse effect on the VM for some reason. Now you do it in stages. Reduce by half, monitor the application and if all is still good re run the right size.

[RobertWLSEG]

Perfect, many thanks for the response. Interestingly enough, that is what a lot of my supermetrics were about as the old metrics were too agressive!