What is the potential performance impact of creating a supermetric for all VMs in a datacenter object (4000 vms total) and applying to that datacenter object? Is this risky in terms of creating performance demand on vROps?
Thanks!
The performance impact should be negligible if any. We have several super metrics that are calculated over 1000's of objects. I don't know of any hard and fast rules on this, but with enough super metrics you'll eventually impact the performance of analytics.
There are however metrics for monitoring super metrics
In Environment under your vROps cluster select "vRealize Operations Manager Analytics" and look for Overall threshold checking | Super metric computation | Duration | Average(ms) or Maximum(ms). You should be able to see if the time increases significantly after creating your super metric. Our average is typically under 1ms.
The performance impact should be negligible if any. We have several super metrics that are calculated over 1000's of objects. I don't know of any hard and fast rules on this, but with enough super metrics you'll eventually impact the performance of analytics.
There are however metrics for monitoring super metrics
In Environment under your vROps cluster select "vRealize Operations Manager Analytics" and look for Overall threshold checking | Super metric computation | Duration | Average(ms) or Maximum(ms). You should be able to see if the time increases significantly after creating your super metric. Our average is typically under 1ms.
Thanks again