Hello,
I’m using Rancher in v2.6.9. I have clusters in RKE 1.24 and I deployed monitoring app in v100.2.0+up40.1.2 (the last available).
Since v1.24, I can’t see metrics (CPU/RAM) for PODs in Grafana. I saw that metric “container_cpu_usage_seconds_total” does not have fields “container”, “images” or “name”.
Before v1.24:
container_cpu_usage_seconds_total{container=“POD”, cpu=“total”, endpoint=“https-metrics”, id=“/kubepods/besteffort/pod00838499-f46a-423d-808e-34da7ee0dabb/7bf4b0e61b19fd28d5b36c529cc8b52036a1b33157c380313a20f6054fc32cd6”, image=“rancher/mirrored-pause:3.6”, instance=“X.X:X:X:10250”, job=“kubelet”, metrics_path=“/metrics/cadvisor”, name=“k8s_POD_pushprox-kube-controller-manager-client-5jchz_cattle-monitoring-system_00838499-f46a-423d-808e-34da7ee0dabb_0”, namespace=“cattle-monitoring-system”, node=“*****”, pod=“pushprox-kube-controller-manager-client-5jchz”, service=“rancher-monitoring-kubelet”}
At v1.24:
container_cpu_usage_seconds_total{cpu=“total”, endpoint=“https-metrics”, id=“/kubepods/besteffort/pod2d532b02-19a4-4da6-bb03-76921d537fe6”, instance=“Y.Y.Y.Y:10250”, job=“kubelet”, metrics_path=“/metrics/cadvisor”, namespace=“cattle-monitoring-system”, node=“*****”, pod=“pushprox-kube-controller-manager-client-6s4rc”, service=“rancher-monitoring-kubelet”}
Others metrics have the same issue. So, I think it is due to some changes in v1.24. The big change was about Dockershim.
Any suggestions to solve the issue?
Thanks,
Yannaël