Kube-proxy healthz probe fails despite it's running

k3s, k3OS, and k3d
1

I have a new installation of RKE2 cluster of 1 master node and 1 worker node that’re running rke2 v1.29.3. The problem is that healthcheck of kube-proxy of the node is failing with connection refused

Startup probe failed: Get "http://localhost:10256/healthz": dial tcp [::1]:10256: connect: connection refused

but I can do a curl and get a response from within the container

$ kubectl exec -it -n kube-system  kube-proxy-node -- curl http://localhost:10256/healthz
{"lastUpdated": "2024-04-29 07:24:26.951815018 +0000 UTC m=+231.679213685","currentTime": "2024-04-29 07:24:26.951815018 +0000 UTC m=+231.679213685", "nodeEligible": true}

I tried uninstalling then reinstalling rke2-agent without success. I also made sure that after uninstallation, there’re no iptables rules leftovers.

The problem doesnt appear on master node