windows node daemonset should have a priority class

[lmm]

Expected Behavior

Critical component resources like the windows-upgrader daemonset should have a priority class.

Current Behavior

The windows-upgrader daemonset does not have a priority class.

Possible Solution

Add a priority class to the windows-upgrader daemonset template spec.

Steps to Reproduce (for bugs)

1. Run an AKS cluster with calico (v3.21.4) policy enabled and windows nodes
2. Observe that the windows-upgrader daemonset doesn't have a priority class

Context

This was raised by the AKS team. Not having a priority class means that the windows-upgrader pods can be preempted and disrupt the upgrade process.

Your Environment

• Operating System and version:
• Link to your project (optional):

[tmjd]

@coutinhop WDYT if I rename this issue to be for the windows daemonset (calico-node-windows) because it has replaced the upgrader daemonset and it seems like the newer daemonset should have the Node Critical priority class? Is there any reason the calico-node-windows daemonset shouldn't have that priority class?

[coutinhop]

@tmjd agreed, the missing priority class was probably and oversight, will add