This message was deleted.

To pre-emptive strike any questions about it, I do not have --no-flannel or --disable flags more than for traefik, and I have --flannel-backend wireguard-native flags set in the systemd file.

If you look on the "How it works" picture on here https://k3s.io/ you should see that Flannel is part of the k3s process, along with kube proxy kubelet etc. You should not see any of those components as daemonsets or pods. I can only guess that you inadvertently added flannel as a daemonset yourself some time in the few years you've been running the cluster, thinking that you were updating the service, when you were actually just wasting CPU/memory/carbon. Does that sound possible?

Yes, absolutely, and that explains a lot, thank you very much for the response! :)