Explore using the resilient creation/deletion features for VMSS to improve reliability of Helix VMs

[riarenas]

The VMSS team reached out to us to showcase the resilient VM creation/deletion feature. This feature ensures that all VM instance creation is retried up to five times. This should be helpful when dealing with OS provisioning timeouts and other infrastructure problems on the azure side.

Release Note Category

• Feature changes/additions
• Bug fixes
• Internal Infrastructure Improvements

Release Note Description

[riarenas]

This feature has been enabled in the windows.11.arm64.open queue which our telemetry shows sees a lot of machines that never get to heartbeat and are afterwards cleaned up by the dead vm cleaner.

We will monitor the data to see if this creates any improvement over our current VM deletion process.