It’s definitely a great query to have. You would like to see which pods are running close to their limits percentage-wise, right? Then you would have an instant view of all pods and where to adjust.

I tried it shortly and were able to use it in the example app of this article (Kubernetes 1.16) with:

sum(rate(container_cpu_usage_seconds_total{name!~".*prometheus.*", image!="", container!="POD"}[5m])) by (pod, container) /
sum(container_spec_cpu_quota{name!~".*prometheus.*", image!="", container!="POD"}/container_spec_cpu_period{name!~".*prometheus.*", image!="", container!="POD"}) by (pod, container) * 100

I think this query shows all pods CPU usage in percent related to a limit set. I got it from here https://github.com/google/cadvisor/issues/2026

But I didn’t test it thoroughly. Let me know if you got that kind of query running and feel free to post it here :)