Docker CPU reporting > 100% ?

[Jeppedy]

I have Docker containers reporting > 100% utilization during times of heavy use.
First, are the interwebs right that a 3-CPU VM config looks like 300% capacity in the Docker Container?
Second, if that's true, what's best practice for setting alert levels? I often tune VM configs based on what's running on them, as workloads ebb and flow.

[VGBLTSK]

Same here. Not sure how to interprete it.

[rcourtman]

I fixed this on pulse/v6-release in b1fdefdd5. Docker and Podman container CPU now keeps the raw runtime per-core CPU value as diagnostic metadata, but charts, canonical resource metrics, metrics-history fallback, and container CPU alerts use host-capacity-normalized percent. That means a container using 240% on a 4-core host is evaluated as 60% capacity, so normal multi-core Docker usage should no longer trip thresholds as if it were over 100%.