I am working on adding e2e tests to detect when a node goes not ready due to unexpected reasons.

I added new intervals to help detect this.

I have discovered that MachineSetScaling test scales up a node correctly and then the node is scaled down. On the scale down, the monitor tests are detecting that the node is not ready forever.

I would expect that the node is deleted and the status should not be unknown.

Slack Thread: https://redhat-internal.slack.com/archives/CBZHF4DHC/p1723581253794909

https://sippy.dptools.openshift.org/sippy-ng/job_runs/1823458658017611776/pull-ci-openshift-origin-master-e2e-aws-ovn-serial/openshift_origin/28989/intervals?filterText=&intervalFile=e2e-events_20240813-213424.json&overrideDisplayFlag=0&selectedSources=OperatorAvailable&selectedSources=E2EFailed&selectedSources=NodeUnreachable&selectedSources=NodeUnexpectedNotReady&selectedSources=E2EPassed&selectedSources=NodeState

You can see this with the NodeState intevals and see that the Node goes not ready and stays not ready when the machine set scaling test runs.

Could someone look into this from the machine set side?