update cnpg config

carpenike · Nov 25, 2024 · f21885e · f21885e
1 parent c97c926
commit f21885e
Show file tree

Hide file tree

Showing 5 changed files with 47 additions and 36 deletions.
diff --git a/kubernetes/cluster-0/apps/db/cloudnative-pg/app/helmrelease.yaml b/kubernetes/cluster-0/apps/db/cloudnative-pg/app/helmrelease.yaml
@@ -1,14 +1,15 @@
 ---
-apiVersion: helm.toolkit.fluxcd.io/v2beta2
+# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/helm.toolkit.fluxcd.io/helmrelease_v2.json
+apiVersion: helm.toolkit.fluxcd.io/v2
 kind: HelmRelease
 metadata:
   name: cloudnative-pg
 spec:
-  interval: 15m
+  interval: 30m
   chart:
     spec:
       chart: cloudnative-pg
-      version: 0.21.5
+      version: 0.22.1
       sourceRef:
         kind: HelmRepository
         name: cloudnative-pg
@@ -21,8 +22,6 @@ spec:
     remediation:
       strategy: rollback
       retries: 3
-  uninstall:
-    keepHistory: false
   values:
     crds:
       create: true

diff --git a/kubernetes/cluster-0/apps/db/cloudnative-pg/cluster/cluster.yaml b/kubernetes/cluster-0/apps/db/cloudnative-pg/cluster/cluster.yaml
@@ -13,20 +13,23 @@ spec:
     storageClass: local-hostpath
   superuserSecret:
     name: cloudnative-pg-secret
-  bootstrap:
-    recovery:
-      source: &previous-cluster postgres-v8
+  enableSuperuserAccess: true
   postgresql:
     parameters:
       max_connections: "300"
       shared_buffers: 512MB
       pg_stat_statements.max: "10000"
       pg_stat_statements.track: all
+  resources:
+    requests:
+      cpu: 500m
+    limits:
+      memory: 4Gi
   monitoring:
     enablePodMonitor: true
   backup:
     retentionPolicy: 30d
-    barmanObjectStore:
+    barmanObjectStore: &barmanObjectStore
       wal:
         compression: bzip2
         maxParallel: 8
@@ -40,19 +43,13 @@ spec:
         secretAccessKey:
           name: cloudnative-pg-secret
           key: aws-secret-access-key
+  bootstrap:
+    recovery:
+      source: &previousCluster postgres-v8
+
   # Note: externalClusters is needed when recovering from an existing cnpg cluster
   externalClusters:
-    - name: *previous-cluster
+    - name: *previousCluster
       barmanObjectStore:
-        wal:
-          compression: bzip2
-          maxParallel: 8
-        destinationPath: s3://postgresql/
-        endpointURL: https://minio-s3.holthome.net
-        s3Credentials:
-          accessKeyId:
-            name: cloudnative-pg-secret
-            key: aws-access-key-id
-          secretAccessKey:
-            name: cloudnative-pg-secret
-            key: aws-secret-access-key
+        <<: *barmanObjectStore
+        serverName: *previousCluster
diff --git a/kubernetes/cluster-0/apps/db/cloudnative-pg/cluster/prometheusrule.yaml b/kubernetes/cluster-0/apps/db/cloudnative-pg/cluster/prometheusrule.yaml
@@ -1,5 +1,5 @@
 ---
-# yaml-language-server: $schema=https://kubernetes-schemas.devbu.io/monitoring.coreos.com/prometheusrule_v1.json
+# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/monitoring.coreos.com/prometheusrule_v1.json
 apiVersion: monitoring.coreos.com/v1
 kind: PrometheusRule
 metadata:
@@ -31,10 +31,10 @@ spec:
             severity: warning
         - alert: PGDatabase
           annotations:
-            description: Over 150,000,000 transactions from frozen xid on pod {{ $labels.pod }}
+            description: Over 300,000,000 transactions from frozen xid on pod {{ $labels.pod }}
             summary: Number of transactions from the frozen XID to the current one
           expr: |-
-            cnpg_pg_database_xid_age > 150000000
+            cnpg_pg_database_xid_age > 300000000
           for: 1m
           labels:
             severity: warning
@@ -50,9 +50,9 @@ spec:
         - alert: LastFailedArchiveTime
           annotations:
             description: Archiving failed for {{ $labels.pod }}
-            summary: Checks the last time archiving failed. Will be -1 when it has not failed.
+            summary: Checks the last time archiving failed. Will be < 0 when it has not failed.
           expr: |-
-            delta(cnpg_pg_stat_archiver_last_failed_time[5m]) > 0
+            (cnpg_pg_stat_archiver_last_failed_time - cnpg_pg_stat_archiver_last_archived_time) > 1
           for: 1m
           labels:
             severity: warning
@@ -65,3 +65,12 @@ spec:
           for: 1m
           labels:
             severity: warning
+        - alert: ReplicaFailingReplication
+          annotations:
+            description: Replica {{ $labels.pod }} is failing to replicate
+            summary: Checks if the replica is failing to replicate
+          expr: |-
+            cnpg_pg_replication_in_recovery > cnpg_pg_replication_is_wal_receiver_up
+          for: 1m
+          labels:
+            severity: warning
diff --git a/kubernetes/cluster-0/apps/db/cloudnative-pg/cluster/service.yaml b/kubernetes/cluster-0/apps/db/cloudnative-pg/cluster/service.yaml
@@ -3,7 +3,6 @@ apiVersion: v1
 kind: Service
 metadata:
   name: postgres-lb
-  namespace: db
   annotations:
     external-dns.alpha.kubernetes.io/hostname: postgres.holthome.net
     io.cilium/lb-ipam-ips: 10.45.20.15
@@ -16,4 +15,4 @@ spec:
       targetPort: 5432
   selector:
     cnpg.io/cluster: postgres
-    role: primary
+    cnpg.io/instanceRole: primary
diff --git a/kubernetes/cluster-0/apps/db/cloudnative-pg/ks.yaml b/kubernetes/cluster-0/apps/db/cloudnative-pg/ks.yaml
@@ -1,37 +1,44 @@
 ---
+# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/kustomize.toolkit.fluxcd.io/kustomization_v1.json
 apiVersion: kustomize.toolkit.fluxcd.io/v1
 kind: Kustomization
 metadata:
-  name: cloudnative-pg
+  name: &app cloudnative-pg
   namespace: flux-system
 spec:
   targetNamespace: db
+  commonMetadata:
+    labels:
+      app.kubernetes.io/name: *app
   path: ./kubernetes/cluster-0/apps/db/cloudnative-pg/app
   prune: true
   sourceRef:
     kind: GitRepository
     name: k8s-gitops-kubernetes
   wait: true
   interval: 30m
-  retryInterval: 1m
   timeout: 5m
+  dependsOn:
+    - name: external-secrets-stores
 ---
+# yaml-language-server: $schema=https://kubernetes-schemas.pages.dev/kustomize.toolkit.fluxcd.io/kustomization_v1.json
 apiVersion: kustomize.toolkit.fluxcd.io/v1
 kind: Kustomization
 metadata:
-  name: cloudnative-pg-cluster
+  name: &app cloudnative-pg-cluster
   namespace: flux-system
 spec:
   targetNamespace: db
-  dependsOn:
-    - name: cloudnative-pg
-    - name: external-secrets-stores
+  commonMetadata:
+    labels:
+      app.kubernetes.io/name: *app
   path: ./kubernetes/cluster-0/apps/db/cloudnative-pg/cluster
   prune: true
   sourceRef:
     kind: GitRepository
     name: k8s-gitops-kubernetes
   wait: true
   interval: 30m
-  retryInterval: 1m
   timeout: 5m
+  dependsOn:
+    - name: cloudnative-pg