This message was deleted.
# k3s
a
This message was deleted.
h
you can start with looking at resource utilization of pods across all namespaces:
kubectl top po -A
and looking at the logs for k3s?
journalctl -u k3s-server.service
or it may be just
journalctl -u k3s.service
look at the release notes to see if this is a known issue for the version you have running?
b
regarding
kubectl top po
Its not the pods who consume the cpu, but the k3s server process itself
The main "k3s server" process has a lot of threads. I know that some them are: • api-server • scheduler • kind (sqlite wrapped as etc) but I have 113 threads. How can I figure out what each thread does?
h
I no long run k3s so from k3s perspective I don't know... from Linux side you can check what each process is doing with lsof
n
Coming to this late, but you can watch all the pids associated with the K3s cgroup with
systemd-cgtop system.slice/k3s.service