VM's with high CPU Ready

[sidkipper]

(I'm new to Reporter, so have probably missed something obvious, apologies in advance)

I've been trying to generate a report showing all VM's that have been experiencing high CPU ready times (say above 1000ms) for a particular time period, without much luck, and suspect I'm going about it the wrong way.

I've generated various reports using the following as a base, but I always end up with a graph loads of VM's and its impossible to tell which of those are actually suffering.

• Report Pack: Performance report pack
  Report: VM Performance (template)
  Performance Counter:CPU Read

Is there a way to generate a table rather than a graph, with a filter to exclude VM's with low CPU Ready times? Or another suggested method?

[Vitaliy S.]

Hello Simon,

Yes, I agree that it might be difficult to figure out what VMs are actually suffering from high CPU Ready, that is why we are introducing a new, more detailed report pack in the next release which is scheduled for Q4.

Meanwhile, if you're interested in detecting VMs with high CPU Ready, you can deploy Veeam Monitor and run a corresponding report from there. Free version of Veeam Monitor is also capable to do that.

By the way, do you have any other reports you would like to see in Veeam Reporter?

Thanks.

[sidkipper]

Thanks - nice to know I wasn't being blind!

From what I've tried to do so far I think the capacity/resource management side of things could be beefed up, for example...

Being able to detect over-provisioned VM's (too much/many vCPU's, memory, thick disk space)
- EG A report might show VM's with multiple vCPU's, that never use more than say 30% CPU

In the Capacity Planning report pack it would be nice to be able to specify "Resource utilization threshold" for clusters in "n+1", "n+2" etc style, for when you want all your clusters to have 1, 2, etc ESX's of fail-over capacity available. This gets a bit more complicated as even in a fail-over scenario, you don't really want your surviving ESX's to be using more than around 90% memory and 75% CPU (CPU especially as you tend to get the equivalent of thermal runaway when ESX's are heavily CPU laden).

On a different more niche note, being able to report on ESX hardware NIC and HBA driver version would be useful. I suspect this could be a pain to do in a reliable and uniform fashion as you're dependant on the info being provided by the server manufacturer's CIM implementation.

If there's anything else that comes to mind I'll let you know!

Cheers
Simon

[Vitaliy S.]

Being able to detect over-provisioned VM's (too much/many vCPU's, memory, thick disk space)

Already implemented!

In the Capacity Planning report pack it would be nice to be able to specify "Resource utilization threshold" for clusters in "n+1", "n+2" etc style, for when you want all your clusters to have 1, 2, etc ESX's of fail-over capacity available. This gets a bit more complicated as even in a fail-over scenario, you don't really want your surviving ESX's to be using more than around 90% memory and 75% CPU (CPU especially as you tend to get the equivalent of thermal runaway when ESX's are heavily CPU laden).

Got it, thanks for the feedback.

On a different more niche note, being able to report on ESX hardware NIC and HBA driver version would be useful. I suspect this could be a pain to do in a reliable and uniform fashion as you're dependant on the info being provided by the server manufacturer's CIM implementation.

Hmm...let me see if this information can be retrieved easily.

[howthorn]

I would appreciate if you make a brief description how to report over-provisioned VMs, either using monitor or reporter.

thanks...

[Vitaliy S.]

Hi Ilkay,

Currently, there is no easy way to create such report, however it will be available in our next version (v6).

Thanks.