I did an RKE2 upgrade recently from 1.25.10 to 1.26.10+rke2r2. After the upgrade I had several network issue on my workers. After some debugging I found that the kube-proxy pods were not restarted on the workers. Is this a known issue? So whoever is experiencing this: Don't mess with kube-proxy manually or kubelet restart, it makes things worse. Simply reboot the workers 🙂