mistake metric value of kubernetes.pod.cpu.usage.limit.pct

[zhongxp1219]

Hi, We are using metricbeat to collects k8s pods cpu. In the beginning, it can collects the cpu correctly. After running for serveral weeks or serveral months, we found that one of the pods cpu usage metricbeat collects is lower than actually used. Finally, it recovers while we restart the metricbeat. We can see as follows graph.

1.Metricbeat configuration:

2.Cpu usage in grafana:

By the way, the node is 16U, and i found that the code will use nodeCore to calculate if failed to find container cores

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

Line 129 in a55ae58

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

the request and limit of service is 1U, the node is 16U, if nodeCores was used to calculate, then the cpu.usage is equals to %3.
the problem is that may be the nodeCores is used to calculate for kubernetes.pod.cpu.usage.limit.pct ?

For confirmed bugs, please report:

Version: 7.6.0
Operating System: EulerOS 2.7
Discuss Forum URL: https://discuss.elastic.co/t/metricbeat-collects-k8s-pods-cpu-which-is-lower-than-actually-used/279215
Steps to Reproduce: it happenes randomly

[botelastic]

This issue doesn't have a Team:<team> label.

[zhongxp1219]

hi，anybody can help to fix it ？

[zhongxp1219]

hi，anybody can help to fix it ？

[botelastic]

Hi!
We just realized that we haven't looked into this issue in a while. We're sorry!

We're labeling this issue as Stale to make it hit our filters and make sure we get back to it as soon as possible. In the meantime, it'd be extremely helpful if you could take a look at it as well and confirm its relevance. A simple comment with a nice emoji will be enough :+1.
Thank you for your contribution!