This message was deleted.

Are you using authorised cluster endpoint / fronting these nodes with a LB? running something like

kubectl -v3 get po

can potentially highlight what's going on

That command showed an issue with the metrics server. Was able to find an issue on on the rke2 GitHub with someone having the exact same issue https://github.com/rancher/rke2/issues/745 Applyin the network policy Brandon suggested fixed the issue

Interesting, thanks for the update