split ksm alerts in 2 separate ones (#912)

* split ksm alerts in 2 separate ones * move new alert to the adequate file * moved KSMDown alert to the adequate file * fix rules * fix chart * minor fixes * lowered down time to trigger to 20m * changelog
giantswarm · Sep 21, 2023 · 47fcf66 · 47fcf66
1 parent 42721e2
commit 47fcf66
Show file tree

Hide file tree

Showing 4 changed files with 54 additions and 34 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Changed
+
+- Split `KubeStateMetricsDown` alert into 2 alerts : `KubeStateMetricsDown` and `KubeStateMetricsNotRetrievingMetrics`
+
 ## [2.133.0] - 2023-09-19
 
 ### Changed

diff --git a/helm/prometheus-rules/templates/alerting-rules/kube-state-metrics.rules.yml b/helm/prometheus-rules/templates/alerting-rules/kube-state-metrics.rules.yml
@@ -10,6 +10,34 @@ spec:
   groups:
   - name: kube-state-metrics
     rules:
+    - alert: KubeStateMetricsDown
+      annotations:
+        description: '{{`KubeStateMetrics ({{ $labels.instance }}) is down.`}}'
+        opsrecipe: kube-state-metrics-down/
+      expr: |-
+        (
+          # modern clusters
+          label_replace(up{app="kube-state-metrics",instance=~".*:8080"}, "ip", "$1.$2.$3.$4", "node", "ip-(\\d+)-(\\d+)-(\\d+)-(\\d+).*") == 0 or absent(up{app="kube-state-metrics",instance=~".*:8080"} == 1)
+        )
+        and
+        (
+          # vintage clusters without servicemonitor
+          label_replace(up{app="kube-state-metrics",container=""}, "ip", "$1.$2.$3.$4", "node", "ip-(\\d+)-(\\d+)-(\\d+)-(\\d+).*") == 0 or absent(up{app="kube-state-metrics",container=""} == 1)
+        )
+      for: 15m
+      labels:
+        area: kaas
+        cancel_if_apiserver_down: "true"
+        cancel_if_cluster_status_creating: "true"
+        cancel_if_cluster_status_deleting: "true"
+        cancel_if_cluster_has_no_workers: "true"
+        inhibit_kube_state_metrics_down: "true"
+        cancel_if_prometheus_agent_down: "true"
+        cancel_if_kubelet_down: "true"
+        cancel_if_outside_working_hours: "false"
+        severity: page
+        team: atlas
+        topic: observability
     - alert: KubeStateMetricsSlow
       annotations:
         description: '{{`KubeStateMetrics ({{ $labels.instance }}) is too slow.`}}'
@@ -28,6 +56,27 @@ spec:
         severity: page
         team: atlas
         topic: observability
+    - alert: KubeStateMetricsNotRetrievingMetrics
+      annotations:
+        description: '{{`KubeStateMetrics ({{ $labels.instance }}) is not retrieving metrics.`}}'
+        opsrecipe: kube-state-metrics-down/
+      expr: |-
+        # When it looks up but we don't have metrics
+        count({app="kube-state-metrics"}) < 10
+      for: 20m
+      labels:
+        area: kaas
+        cancel_if_apiserver_down: "true"
+        cancel_if_cluster_status_creating: "true"
+        cancel_if_cluster_status_deleting: "true"
+        cancel_if_cluster_has_no_workers: "true"
+        inhibit_kube_state_metrics_down: "true"
+        cancel_if_kubelet_down: "true"
+        cancel_if_kube_state_metrics_down: "true"
+        cancel_if_outside_working_hours: "true"
+        severity: page
+        team: atlas
+        topic: observability
     - alert: KubeConfigMapCreatedMetricMissing
       annotations:
         description: '{{`kube_configmap_created metric is missing for cluster {{ $labels.cluster_id }}.`}}'

diff --git a/helm/prometheus-rules/templates/alerting-rules/up.all.rules.yml b/helm/prometheus-rules/templates/alerting-rules/up.all.rules.yml
@@ -46,36 +46,3 @@ spec:
         severity: page
         team: atlas
         topic: observability
-    - alert: KubeStateMetricsDown
-      annotations:
-        description: '{{`KubeStateMetrics ({{ $labels.instance }}) is down.`}}'
-        opsrecipe: kube-state-metrics-down/
-      expr: |-
-        (
-          # modern clusters
-          label_replace(up{app="kube-state-metrics",instance=~".*:8080"}, "ip", "$1.$2.$3.$4", "node", "ip-(\\d+)-(\\d+)-(\\d+)-(\\d+).*") == 0 or absent(up{app="kube-state-metrics",instance=~".*:8080"} == 1)
-        )
-        and
-        (
-          # vintage clusters without servicemonitor
-          label_replace(up{app="kube-state-metrics",container=""}, "ip", "$1.$2.$3.$4", "node", "ip-(\\d+)-(\\d+)-(\\d+)-(\\d+).*") == 0 or absent(up{app="kube-state-metrics",container=""} == 1)
-        )
-        or
-        (
-          # When it looks up but we don't have metrics
-          count({app="kube-state-metrics"}) < 10
-        )
-      for: 15m
-      labels:
-        area: kaas
-        cancel_if_apiserver_down: "true"
-        cancel_if_cluster_status_creating: "true"
-        cancel_if_cluster_status_deleting: "true"
-        cancel_if_cluster_has_no_workers: "true"
-        inhibit_kube_state_metrics_down: "true"
-        cancel_if_kubelet_down: "true"
-        cancel_if_outside_working_hours: "false"
-        cancel_if_prometheus_agent_down: "true"
-        severity: page
-        team: atlas
-        topic: observability
diff --git a/...ts/providers/global/up.all.rules.test.yml → .../global/kube-state-metrics.rules.test.yml b/...ts/providers/global/up.all.rules.test.yml → .../global/kube-state-metrics.rules.test.yml
@@ -1,6 +1,6 @@
 ---
 rule_files:
-- up.all.rules.yml
+- kube-state-metrics.rules.yml
 
 tests:
   # KubeStateMetricsDown tests