charts,docs,salt : Add an argument to kube-state-metrics container to exclude selected PVCs from the metrics

Author: ezekiel-alexrod

CHANGELOG.md

@@ -7,6 +7,14 @@
 - Do not deploy Loki ingress when Loki addon is not enabled
   (PR[#4725](https://github.com/scality/metalk8s/pull/4725))
 
+### Enhancements
+
+- Add a patch to the `kube-prometheus-stack` chart to add
+  the `--metric-labels-allowlist=persistentvolumeclaims=[excluded-from-alerts]`
+  argument to the `kube-state-metrics` container to exclude PVCs
+  from the `kube-state-metrics` metrics.
+  (PR[#4723](https://github.com/scality/metalk8s/pull/4723))
+
 ## Release 131.0.5
 
 ## Release 131.0.4

charts/kube-prometheus-stack.yaml

@@ -246,6 +246,9 @@ kube-state-metrics:
     registry: '__var__(repo.registry_endpoint)'
     repository: '__image_no_reg__(kube-state-metrics)'
 
+  metricLabelsAllowlist:
+    - persistentvolumeclaims=[excluded-from-alerts]
+
   nodeSelector:
     node-role.kubernetes.io/infra: ''
 

docs/operation/cluster_monitoring/alerting.rst

@@ -40,3 +40,39 @@ Simple Rules
 .. jinja:: alerting
    :file: _jinja/alerting/simple-rulelist.rst.j2
    :header_char: "
+
+Excluding PersistentVolumeClaims from Storage Alerts
+-----------------------------------------------------
+
+MetalK8s monitors PersistentVolumeClaim (PVC) storage usage and generates
+alerts when volumes are filling up. To exclude specific PVCs from these
+storage-related alerts (such as ``KubePersistentVolumeFillingUp`` and
+``KubePersistentVolumeInodesFillingUp``), add the ``excluded-from-alerts``
+label with the value ``"true"`` to the PVC.
+
+.. code-block:: yaml
+
+   ---
+   apiVersion: v1
+   kind: PersistentVolumeClaim
+   metadata:
+     name: <pvc-name>
+     namespace: <namespace-name>
+     labels:
+       excluded-from-alerts: "true"
+   spec:
+     # ... PVC specification ...
+
+To add this label to an existing PVC:
+
+.. code-block:: shell
+
+    root@bootstrap $ kubectl --kubeconfig=/etc/kubernetes/admin.conf \
+                       label pvc <pvc-name> -n <namespace-name> \
+                       excluded-from-alerts=true
+
+.. note::
+
+   This feature is particularly useful for data disks that are expected to
+   be nearly full or for volumes where storage alerts are managed by
+   application-specific monitoring.

salt/metalk8s/addons/prometheus-operator/deployed/chart.sls

@@ -71798,6 +71798,7 @@ spec:
       - args:
         - --port=8080
         - --resources=certificatesigningrequests,configmaps,cronjobs,daemonsets,deployments,endpoints,horizontalpodautoscalers,ingresses,jobs,leases,limitranges,mutatingwebhookconfigurations,namespaces,networkpolicies,nodes,persistentvolumeclaims,persistentvolumes,poddisruptionbudgets,pods,replicasets,replicationcontrollers,resourcequotas,secrets,services,statefulsets,storageclasses,validatingwebhookconfigurations,volumeattachments
+        - --metric-labels-allowlist=persistentvolumeclaims=[excluded-from-alerts]
         image: {% endraw -%}{{ repo.registry_endpoint }}{%- raw %}/{% endraw -%}{{ build_image_name("kube-state-metrics", False, False) }}{%- raw %}:v2.13.0
         imagePullPolicy: IfNotPresent
         livenessProbe: