Merge pull request prometheus-community#45 from junotx/rules

fix node:node_pod_utilisation:ratio rule
junotx · Feb 27, 2024 · 56a3e06 · 56a3e06
2 parents 86d31a2 + baa65a9
commit 56a3e06
Show file tree

Hide file tree

Showing 4 changed files with 37 additions and 4 deletions.
diff --git a/charts/kube-prometheus-stack/Chart.yaml b/charts/kube-prometheus-stack/Chart.yaml
@@ -21,7 +21,7 @@ name: kube-prometheus-stack
 sources:
   - https://github.com/prometheus-community/helm-charts
   - https://github.com/prometheus-operator/kube-prometheus
-version: 52.1.11
+version: 52.1.12
 appVersion: v0.68.0
 kubeVersion: ">=1.19.0-0"
 home: https://github.com/prometheus-operator/kube-prometheus

diff --git a/...s/kube-prometheus-stack/templates/prometheus/rules-1.14/whizard-telemetry-node.rules.yaml b/...s/kube-prometheus-stack/templates/prometheus/rules-1.14/whizard-telemetry-node.rules.yaml
@@ -234,7 +234,7 @@ spec:
           {{- toYaml . | nindent 8 }}
         {{- end }}
       {{- end }}
-    - expr: node:node_pod_total:sum / node:node_pod_quota:sum
+    - expr: node:node_pod_running_total:sum / node:node_pod_quota:sum
       record: node:node_pod_utilisation:ratio
       {{- if or .Values.defaultRules.additionalRuleLabels .Values.defaultRules.additionalRuleGroupLabels.whizardTelemetry }}
       labels:
@@ -245,6 +245,22 @@ spec:
           {{- toYaml . | nindent 8 }}
         {{- end }}
       {{- end }}
+    - expr: |-
+        count by(cluster, node) (
+            node_namespace_pod:kube_pod_info:
+            unless on (cluster, namespace, pod)
+            (kube_pod_status_phase{job="kube-state-metrics", phase=~"Failed|Pending|Unknown|Succeeded"} > 0)
+        )
+      record: node:node_pod_running_total:sum
+      {{- if or .Values.defaultRules.additionalRuleLabels .Values.defaultRules.additionalRuleGroupLabels.whizardTelemetry }}
+      labels:
+        {{- with .Values.defaultRules.additionalRuleLabels }}
+          {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.defaultRules.additionalRuleGroupLabels.whizardTelemetry }}
+          {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
     - expr: |-
         sum by (cluster, node) (
             sum by (cluster, namespace, pod) (kube_pod_status_scheduled{job="kube-state-metrics"} > 0)

diff --git a/ks-prometheus/components/whizard-telemetry-mixin/rules/custom.libsonnet b/ks-prometheus/components/whizard-telemetry-mixin/rules/custom.libsonnet
@@ -205,7 +205,17 @@
           {
             record: 'node:node_pod_utilisation:ratio',
             expr: |||
-              node:node_pod_total:sum / node:node_pod_quota:sum
+              node:node_pod_running_total:sum / node:node_pod_quota:sum
+            ||| % $._config,
+          },
+          {
+            record: 'node:node_pod_running_total:sum',
+            expr: |||
+              count by(%(clusterLabel)s, node) (
+                  node_namespace_pod:kube_pod_info:
+                  unless on (%(clusterLabel)s, namespace, %(podLabel)s)
+                  (kube_pod_status_phase{%(kubeStateMetricsSelector)s, phase=~"Failed|Pending|Unknown|Succeeded"} > 0)
+              )
             ||| % $._config,
           },   
           {

diff --git a/ks-prometheus/manifests/whizard-telemetry-prometheusRule.yaml b/ks-prometheus/manifests/whizard-telemetry-prometheusRule.yaml
@@ -146,8 +146,15 @@ spec:
         sum by (cluster, node)(node:node_device_filesystem_bytes_total:sum)
       record: node:node_filesystem_bytes_total:sum
     - expr: |
-        node:node_pod_total:sum / node:node_pod_quota:sum
+        node:node_pod_running_total:sum / node:node_pod_quota:sum
       record: node:node_pod_utilisation:ratio
+    - expr: |
+        count by(cluster, node) (
+            node_namespace_pod:kube_pod_info:
+            unless on (cluster, namespace, pod)
+            (kube_pod_status_phase{job="kube-state-metrics", phase=~"Failed|Pending|Unknown|Succeeded"} > 0)
+        )
+      record: node:node_pod_running_total:sum
     - expr: |
         sum by (cluster, node) (
             sum by (cluster, namespace, pod) (kube_pod_status_scheduled{job="kube-state-metrics"} > 0)