Cherry-pick #16128 to 7.5: Improve kubernetes.pod.cpu.usage.limit.pct field description

[ChrsMark]

Cherry-pick of PR #16128 to 7.5 branch. Original message:

What does this PR do?

This PR improves kubernetes.pod.cpu.usage.limit.pct field description to make clear how the pct is calculated when at least one container of a Pod has no limits.

Why is it important?

When at least one container of a Pod has no limits, the situation can be tricky since in such case the limit.pct will fall-back to node.pct.

Example:

apiVersion: v1
kind: Pod
metadata:
  name: cpu-demo3
  namespace: beats
spec:
  containers:
  - name: cpu-demo-ctr
    image: vish/stress
    resources:
      limits:
        cpu: "1"
      requests:
        cpu: "0.5"
    args:
    - -cpus
    - "2"
  - name: cpu-demo-ctr2
    image: vish/stress
    args:
    - -cpus
    - "1"

In this case the sum of CoresLimit calculated at

beats/metricbeat/module/kubernetes/pod/data.go

Line 56 in db51085

coresLimit += perfMetrics.ContainerCoresLimit.GetWithDefault(cuid, nodeCores)

will be greater than nodeCores, cause cpu-demo-ctr2 has no limit and hence it will add nodeCores to the sum.
Given this fact, later on at

beats/metricbeat/module/kubernetes/pod/data.go

Line 109 in db51085

if coresLimit > nodeCores {

the coresLimit will become equal to nodeCores making cpu.usage.node.pct and cpu.usage.limit.pct to be equal as well.

Related issues

• Relates Pod and container usage percentage is not properly calculated #10658

cc @jsoriano @exekias