Alert user if ceph metadata server is consuming cpu at threshold point.

Signed-off-by: Manish <[email protected]>

metrics/mixin/alerts/alerts.libsonnet

@@ -3,4 +3,5 @@
 (import 'services.libsonnet') +
 (import 'blocklist.libsonnet') +
 (import 'encryption.libsonnet') +
-(import 'storage-client.libsonnet')
+(import 'storage-client.libsonnet') +
+(import 'ceph-overload.libsonnet')

metrics/mixin/alerts/perf.libsonnet

@@ -0,0 +1,26 @@
+{
+  prometheusAlerts+:: {
+    groups+: [
+      {
+        name: 'ODF-ceph-mds-high-cpu-warnings.rules',
+        rules: [
+          {
+            alert: 'MDS-high-cpu',
+            expr: |||
+              pod:container_cpu_usage:sum{%(mdsSelector)s}/ on(pod) kube_pod_resource_limit{resource='cpu',%(mdsSelector)s} > 0.67
+            ||| % $._config,
+            'for': $._config.mds_cpu_usage_high_threshold_duration,
+            labels: {
+              severity: 'warning',
+            },
+            annotations: {
+              message: 'Ceph metadata server pod ({{ $labels.pod }}) has high cpu usage',
+              description: 'Ceph metadata server pod ({{ $labels.pod }}) has high cpu usage.\nPlease consider increasing the number of active metadata servers,\nit can be done by increasing the number of activeMetadataServers parameter in the StorageCluster CR.',
+              severity_level: 'warning',
+            },
+          },
+        ],
+      },
+    ],
+  },
+}

metrics/mixin/config.libsonnet

@@ -13,6 +13,8 @@
     odfPoolMirroringImageHealthCriticalAlertTime: '10s',
     blockedRBDClientAlertTime: '10s',
     ocsStorageClusterKMSConnectionAlert: '5s',
+    mdsSelector: 'pod=~"rook-ceph-mds.*"',
+    mds_cpu_usage_high_threshold_duration: '6h',
 
     // Constants
     objectStorageType: 'RGW',