Allow users to specify a node pool VM to delete (with or without scaling)

[eponerine]

Title:
Allow users to specify a node pool VM to delete (with or without scaling)

Description:
In simple terms, it would be slick to provide the logic to target a specific VM to kill; whether independently or when scaling down a node pool.

For scaling, perhaps add a new parameter only validates and works if the -count parameter is less-than current count. Something like -targetVM or -targetNode

As an example, let's say you have the following:

• AKS Cluster Name: k8scluster01
• Pool Name: k8scluster01linuxpool01

And in that pool, you have 3 nodes with VM names:

• k8scluster01linuxpool01-AAAAAAA
• k8scluster01linuxpool01-BBBBBBBB
• k8scluster01linuxpool01-CCCCCCC

And assume that k8scluster01linuxpool01-CCCCCCC has some kind of failure that has to be manually fixed by simply deleting the entire VM from the cluster.

If we had the ability to run Set-AksHciNodePool -clusterName "k8scluster01" -name "k8scluster01linuxpool01" -count 2 -targetNode "k8scluster01linuxpool01-CCCCCCC" , it would be smart enough to see that -count 2 is less than current value and we are targeting a specific VM to remove.

For non-scaling situations where you just want to kill a VM and have it respawn, make a new cmdlet called Remove-AksHciNodePoolVM or Remove-AksHciNodePoolNode that handles this logic.

[eponerine]

This would be useful for this very very old issue:
#8