Merge pull request #2309 from sp98/mds-alerts

add alert for mds cache over use

metrics/deploy/prometheus-ocs-rules.yaml

@@ -265,3 +265,15 @@ spec:
         floor((ocs_storage_provider_operator_version>0)/1000) - ignoring(storage_consumer_name) group_right() floor((ocs_storage_client_operator_version>0)/1000) > 1 or floor((ocs_storage_client_operator_version>0)/1000) - ignoring(storage_consumer_name) group_left() floor((ocs_storage_provider_operator_version>0)/1000) >= 1
       labels:
         severity: critical
+  - name: mds-performance-alerts.rules
+    rules:
+    - alert: MDSCacheUsageHigh
+      annotations:
+        description: MDS cache usage for the daemon {{ $labels.ceph_daemon }} has exceeded above 95% of the requested value. Increase the memory request for {{ $labels.ceph_daemon }} pod.
+        message: High MDS cache usage for the daemon {{ $labels.ceph_daemon }}.
+        severity_level: error
+      expr: |
+        ceph_mds_mem_rss / on(ceph_daemon) group_left(job)(label_replace(kube_pod_container_resource_requests{container="mds", resource="memory"}, "ceph_daemon", "mds.$1", "pod", "rook-ceph-mds-(.*)-(.*)") * .5) > .95
+      for: 5m
+      labels:
+        severity: critical

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 'perf.libsonnet')

metrics/mixin/alerts/perf.libsonnet

@@ -0,0 +1,26 @@
+{
+  prometheusAlerts+:: {
+    groups+: [
+      {
+        name: 'mds-performance-alerts.rules',
+        rules: [
+          {
+            alert: 'MDSCacheUsageHigh',
+            expr: |||
+              ceph_mds_mem_rss / on(ceph_daemon) group_left(job)(label_replace(kube_pod_container_resource_requests{container="mds", resource="memory"}, "ceph_daemon", "mds.$1", "pod", "rook-ceph-mds-(.*)-(.*)") * .5) > .95
+            ||| % $._config,
+            'for': $._config.mdsCacheUsageAlertTime,
+            labels: {
+              severity: 'critical',
+            },
+            annotations: {
+              message: 'High MDS cache usage for the daemon {{ $labels.ceph_daemon }}.',
+              description: 'MDS cache usage for the daemon {{ $labels.ceph_daemon }} has exceeded above 95% of the requested value. Increase the memory request for {{ $labels.ceph_daemon }} pod.',
+              severity_level: 'error',
+            },
+          },
+        ],
+      },
+    ],
+  },
+}

metrics/mixin/config.libsonnet

@@ -13,6 +13,7 @@
     odfPoolMirroringImageHealthCriticalAlertTime: '10s',
     blockedRBDClientAlertTime: '10s',
     ocsStorageClusterKMSConnectionAlert: '5s',
+    mdsCacheUsageAlertTime: '5m',
 
     // Constants
     objectStorageType: 'RGW',