See #131612 (comment)

#131612 introduces a race condition, where if a pod is in the middle of syncing, and a new resize is allocated in between the check that a resize is in progress and where the pod status is calculated, the PodResizeInProgress condition is added back unconditionally and not actually properly checked until the next sync loop. The consequences are small and temporary, so considering this a non-blocker for 1.34.

I think this is only really an issue if a resize is reverted.

/sig node
/triage accepted
/priority important-longterm