How do you group your VM's with large VM environments?

My suggestion would be to clone the DataSource for VM Performance and rename it appropriately for what you are looking for; i.e VMware_vCenter_VMPerformance_DevServers and Displayed as VMware VM Performance - DEV Servers

Then you can create a filter that would narrow down only the Dev servers, which I would guess you could do by name with something like ##WILDALIAS## RegExMatch DEV, so it would look for the word "DEV" in every VM name.

That would give you a grouping of Developer servers under VM Performance.

Cloning the DataSource again and creating a filter that gives you the individual VMs would be the next task.  Then you have them grouped, can report on them, and have separate thresholds for them.

Cloning is one option, but a cleaner option may be to group the instances under that DataSource. 

It looks like there are several properties that should get discovered by default for every VM instance:

'auto.config.alternate_guest_os_Name': vmConfig?.alternateGuestName,
'auto.config.annotation'             : vmConfig?.annotation,
'auto.config.firmware'               : vmConfig?.firmware,
'auto.config.guest_os_full_Name'     : vmConfig?.guestFullName,
'auto.config.guest_os_id'            : vmConfig?.guestId,
'auto.config.managed_by'             : vmConfig?.managedBy?.type ?: "false",
'auto.config.modified'               : vmConfig?.modified?.getTime(),
'auto.config.template'               : vmConfig?.template,
'auto.guest.guest_os_family'         : vmGuest?.guestFamily,
'auto.guest.guest_os_full_name'      : vmGuest?.guestFullName,
'auto.guest.guest_os_id'             : vmGuest?.guestId,
'auto.guest.hostname'                : vmGuest?.hostName,
'auto.guest.tools_version'           : vmGuest?.toolsVersion,
'auto.guest.tools_version_status'    : vmGuest?.toolsVersionStatus2,
'auto.resource_pool'                 : vm?.resourcePool?.name,
'auto.resource_pool_full_path'       : resource_pool_array.reverse().join(' -> '),
'auto.snapshot_count'                : vm?.layoutEx?.snapshot?.size(),
'auto.cluster'                       : esxhost?.parent?.name,
'auto.cluster_full_path'             : cluster_path_array.reverse().join(' -> '),
'auto.runtime.host'                  : esxhost?.name

You could group by any of these by simply setting the "Group method" to "Instance Level Property" and then choosing which property to group by.

Alternatively, you could choose "Regular Expression" as the Group method and define each group with its own regular expression:

For example, if your dev VMs started with "D" and prod VMs started with "P", you would do this:

Development="D.*" Production="P.*"

Of course, you can always manually create instance groups and sort the items into them manually.