From: Michael Hofmann on
gitlab.com
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/2375#note_1321...
just to be clear, that will still not take cgroup limits into account, so any
Docker/Kubernetes deployment that just applies "Limits" will still see all
cpus on the host machine. For comparison, this is what CKI needs to do to keep
parallel jobs from exploding on 96 core k8s nodes:
https://gitlab.com/cki-
project/pipeline-definition/-/blob/main/pipeline/functions/general.yml#L113