15642 Support K3s and Rancher RKE2

Depending on your infrastructure, the usage metrics may be unavailable to the check
plugins `kube_cpu` and `kube_memory`.  With this Werk, it may be possible to enable the missing
metrics by using the field `containerdOverride` in the `values.yaml`. Specifically, this issue is
know to affect K3s and Rancher RKE2.

How to diagnose this issue? The issue stems from cAdvisor not reporting which pod the collected
metrics belong to.  You can check the metrics reported by cAdvisor by running the command
```
kubectl exec -it {checkmk-node-collector-container-metrics-ID} -c cadvisor -n {checkmk-monitoring-ns} -- sh -c "wget -qO- http://localhost:8080/metrics | grep 'container_memory_working_set_bytes'"
```
If command returns metrics, but the label `container_label_io_kubernetes_pod_name` is missing, i.e.
the following command returns an empty output
```
kubectl exec -it sol-checkmk-node-collector-container-metrics-pvl87 -c cadvisor -n sol -- sh -c "wget -qO- http://localhost:8080/metrics | grep 'container_memory_working_set_bytes' | grep 'container_label_io_kubernetes_pod_name'"
```
then you are affected by this issue. This means the container runtime is not mounted in the default
location (`/run/containerd/containerd.sock`). You may adjust runtime by using the
`containerdOverride` flag.

close: #22

Change-Id: Icdafa30d8731ea4e6a713a94eb4246e380f54d93

.werks/15642

@@ -0,0 +1,28 @@
+Title: Support K3s and Rancher RKE2
+Class: feature
+Compatible: compat
+Component: node-collector
+Date: 1694671730
+Knowledge: undoc
+Level: 2
+Version: 2.0.0-alpha.1
+
+Depending on your infrastructure, the usage metrics may be unavailable to the check
+plugins `kube_cpu` and `kube_memory`.  With this Werk, it may be possible to enable the missing
+metrics by using the field `containerdOverride` in the `values.yaml`. Specifically, this issue is
+known to affect K3s and Rancher RKE2.
+
+How to diagnose this issue? The issue stems from cAdvisor not reporting which pod the collected
+metrics belong to.  You can check the metrics reported by cAdvisor by running the command
+```
+kubectl exec -it {checkmk-node-collector-container-metrics-ID} -c cadvisor -n {checkmk-monitoring-ns} -- sh -c "wget -qO- http://localhost:8080/metrics | grep 'container_memory_working_set_bytes'"
+```
+If command returns metrics, but the label `container_label_io_kubernetes_pod_name` is missing, i.e.
+the following command returns an empty output
+```
+kubectl exec -it sol-checkmk-node-collector-container-metrics-pvl87 -c cadvisor -n sol -- sh -c "wget -qO- http://localhost:8080/metrics | grep 'container_memory_working_set_bytes' | grep 'container_label_io_kubernetes_pod_name'"
+```
+then you are affected by this issue. This means the container runtime is not mounted in the default
+location (`/run/containerd/containerd.sock`). You may adjust runtime by using the
+`containerdOverride` flag.
+