[Feature] Force Drain and Force Reimage

[kaarthis]

CX ran their workload, it's all done. None of the running pods are important at this point
CX cannot surge any node (no more GPU, or no more capacity, etc.)
CX do not want any node to be deleted (because they may not be able to create a new one after one is deleted, as the available capacity could be reclaimed by others).
CX want everything to be upgraded as fast as possible.

Essentially requesting an upgrade by force draining / Reimaging in place without consideration of surge capacity or Pods.

[bmoore-msft]

I might suggest an edit to the last sentence in the OP...

An API to request an upgrade and immediate reimage without factoring in any cordon/drain/pdb. The node would be immediately reimaged and all nodes could be immediately reimaged in parallel. For example:

POST https://management.azure.com/subscriptions/00000000-0000-0000-0000-000000000000/resourceGroups/rg1/providers/Microsoft.ContainerService/managedClusters/clustername1/forceReimage?api-version=2024-10-01

{
  "ignorePDB": true
}

Could be called on the managedCluster resource or an individual agentPool.

Where the POST body is there primarily as an "are you sure?" due to the disruptive nature of the operation. The purpose as mentioned in the OP is to upgrade as quickly as possible without relinquishing any nodes/vms.