Exclude Unhealthy VMs from Autoscale Task

Avatar
Balachandra, Vino-RBC

Hi,
This feature request is to exclude the unhealthy VMs from Autoscale task. The unhealthy VMs are getting factored into the Autoscale calculation as being Unassigned VMs which in turn is preventing creation of new spare machines. As the unhealthy VMs are not usable by the users, they should be excluded from Autoscale. 

In the below example, there are 2 Unhealthy VMs, and the minimum unassigned machine count is set to 5. As it is considering the 2 Unhealthy VMs as Unassigned, the host pool only has 3 Unassigned VMs available. 

4

Comments (2 comments)

Sort by
0
Avatar
Marek Sokół

Hello, I confirm. For example in my environment having set always one session host active (during evening), all others are already shut down. The one active becomes unavailable, and autoscale is doing nothing about it (not mentioning auto healing actions), but autoscale shout start up any session host to maintain one active available session host (like configured)

Best regards

Marek

0
Avatar
Erik Langeland

In relation to this kind of “Unhealthy VM”, I've also noticed that these machines do not trigger a pool to be noted as “Unhealthy” in the Host Pools view. Please update that logic so that the pool does not show as “Healthy” when it has any VMs in this “Unhealthy VMs” section.

Please sign in to leave a comment.