502 ent - How can vms be identified that are part of a storage array and grouped for analysis

[ServiceOptimi]

I have a storage array that seems to have bad behavior aroun 2am. I am trying to figure out if its backup or antivirus on vms causing a spike in write and read rates to this lun and esx. How can I identify those vms without picking them out manually from vsphere operations  environment tab? I want to run the mteric weather map to replay activity for a window on 6 metrics....would I have to run the replay six times?

[ServiceOptimi]

Any takers??? I dont want to have to select each datastore in question from dashboard environment to see each vm.

[IamTHEvilONE]

I think you have to do this manually right now.  We are working on a grouping feature ... just not sure if it allows you to auto group based on datastore.

[ServiceOptimi]

Thanks for the answer. It was helpful, but not what I was hoping for.

We are troubleshooting a spike in behavior tied to a specific array and there is no realistic way to determine the vms that are associated with that array and then present key metrics in a graph to see which vms are spiking outside the rest.

Will the future grouping option be automatic so as each vm associated with an array (or any mechanism) is discovered it will pop into the key group? Todays options are too error prone to provide clear results and trying to group the vms and create a line graph or sparkline for each vm on a key metric overlaid isnt possible either without jumping through a bunch of hoops. Or am I missing something?

[gravesg]

Looking for a similiar function.

How on earth to you track down the "noisy" VM on a datastore? Datastore view does not breakdown metrics per child object (VM). A simple heatmap would suffice, but its not intuitive to build.

[IamTHEvilONE]

you should be able to do a heatmap where you can group by datastore, size by X, color by noise (or some other metric).

[IamTHEvilONE]

there are 3 pre-defined heatmaps in vCOps, and you can define your own ... I did a poor example, but it works to a point.

[gravesg]

Thanks Jon, your example helped conceptualize how to generate a heat map. Size focus is still a bit unclear but the screenshot gives me the info I need. Only thing is I then have to drill down at the VM level to see which vdisk is noisy which for many of my VMs, they live in different datastores.