feat(zookeeper): decouple metrics and fluentd in zk-exporter

.github/workflows/release-zookeeper-ghcr.yml

@@ -0,0 +1,61 @@
+name: Release Zookeeper Chart to GHCR
+
+# Publishes the zookeeper chart to ghcr.io/radiantlogic-devops/helm/zookeeper-dev
+# as an OCI artifact. This is the dev-channel counterpart to the GitHub Pages
+# helm repository at https://radiantlogic-devops.github.io/helm/ and allows
+# consumers (e.g. helm-v8 fid chart dependency) to pull feature-branch builds
+# directly without waiting for a master merge.
+#
+# Trigger model:
+#   - push to master  -> publishes with the Chart.yaml version
+#   - workflow_dispatch -> can be run manually from any branch to produce a
+#                         testable OCI artifact tagged with that branch's
+#                         Chart.yaml version (overwrites if same version).
+#
+# To test a feature branch:
+#   1. Bump charts/zookeeper/Chart.yaml version (e.g. 0.1.7-my-feature)
+#   2. Push the branch
+#   3. Manually trigger this workflow on that branch
+#   4. helm pull oci://ghcr.io/radiantlogic-devops/helm/zookeeper-dev --version <version>
+
+on:
+  push:
+    branches:
+      - master
+    paths:
+      - 'charts/zookeeper/**'
+      - '.github/workflows/release-zookeeper-ghcr.yml'
+  workflow_dispatch:
+
+jobs:
+  release-ghcr:
+    permissions:
+      contents: read
+      packages: write
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Read zookeeper chart version
+        uses: mikefarah/yq@master
+        id: read_zk_chart
+        with:
+          cmd: yq '.version' charts/zookeeper/Chart.yaml
+
+      - name: Publish zookeeper chart to GHCR
+        uses: appany/helm-oci-chart-releaser@v0.4.1
+        with:
+          name: zookeeper-dev
+          repository: helm
+          tag: ${{ steps.read_zk_chart.outputs.result }}
+          registry: ghcr.io/${{ github.repository_owner }}
+          path: charts/zookeeper
+          registry_username: ${{ github.actor }}
+          registry_password: ${{ secrets.GITHUB_TOKEN }}
+          update_dependencies: 'false'
+
+      # Consumers can pull with:
+      #   helm pull oci://ghcr.io/radiantlogic-devops/helm/zookeeper-dev --version <version>

charts/zookeeper/Chart.yaml

@@ -15,7 +15,7 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.1.6
+version: 0.1.7
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to

charts/zookeeper/templates/service.yaml

@@ -8,6 +8,11 @@ spec:
   type: {{ .Values.service.type }}
 {{- if contains "ClusterIP" .Values.service.type }}
   clusterIP: None
+  # Required for StatefulSet pod-DNS resolution during cluster bootstrap.
+  # ZK pods need DNS for sibling pods (zookeeper-0,1,2) BEFORE they are Ready
+  # so they can form a quorum. Without this, K8s 1.22+ EndpointSlice handling
+  # withholds DNS records until pods are Ready -> chicken-and-egg deadlock.
+  publishNotReadyAddresses: true
 {{- end }}
   ports:
     - port: {{ .Values.service.port }}

charts/zookeeper/templates/statefulset.yaml

@@ -88,7 +88,14 @@ spec:
         command: ["/bin/sh","-c"]
         args: ["/opt/radiantone/run.sh"]
 {{- end }}
-{{- if .Values.metrics.enabled }}
+{{- /*
+  Sidecar container spawns if EITHER metrics.enabled or metrics.fluentd.enabled is true.
+  The two subsystems are independent:
+    - metrics.enabled=true        → zookeeper-exporter Go binary runs (pull-mode at :9095)
+    - metrics.enabled + pushMode  → additionally pushes to Pushgateway via cron
+    - metrics.fluentd.enabled     → Fluentd runs for log forwarding (no dependency on metrics.enabled)
+*/}}
+{{- if or .Values.metrics.enabled .Values.metrics.fluentd.enabled }}
       - name: {{ .Chart.Name }}-exporter
         image: {{ .Values.metrics.image }}:{{ .Values.metrics.imageTag }}
         imagePullPolicy: {{ .Values.image.pullPolicy }}
@@ -99,30 +106,33 @@ spec:
           requests:
             cpu: 100m
             memory: 128Mi
+{{- if .Values.metrics.enabled }}
         ports:
         - containerPort: 9095
           name: exporter
+{{- end }}
         volumeMounts:
         - name: zk-pvc
           mountPath: /opt/radiantone/rli-zookeeper-external
-{{- if eq .Values.metrics.fluentd.enabled true }}
+{{- if .Values.metrics.fluentd.enabled }}
         - name: fluentd-config-volume
           mountPath: /fluentd/etc
 {{- end }}
         securityContext:
           runAsUser: 0
         env:
+{{- if .Values.metrics.enabled }}
         - name: ZK_CONN
           value: {{ .Values.metrics.zk_conn | quote }}
         - name: PUSH_MODE
           value: {{ .Values.metrics.pushMode | quote }}
         - name: PUSHGATEWAY_URI
           value: {{ .Values.metrics.pushGateway | quote }}
-        - name: JOB_NAME
-          value: {{ .Release.Namespace | quote }}
         - name: METRICS_PORT
           value: '9095'
-{{- if hasKey .Values.metrics.fluentd "enabled" }}
+{{- end }}
+        - name: JOB_NAME
+          value: {{ .Release.Namespace | quote }}
 {{- if .Values.metrics.fluentd.enabled }}
         - name: FLUENTD_ENABLE
           value: {{ .Values.metrics.fluentd.enabled | quote }}
@@ -133,7 +143,6 @@ spec:
         - name: ELASTICSEARCH_TYPE
           value: {{ .Values.metrics.fluentd.elasticSearchType | default "elasticsearch" | quote }}
 {{- end }}
-{{- end }}
 {{- end }}
       {{- with .Values.nodeSelector }}
       nodeSelector:

charts/zookeeper/values.yaml

@@ -103,14 +103,26 @@ persistence:
 
 
 metrics:
+  # Controls the zk-exporter sidecar. Two independent subsystems:
+  #   metrics.enabled=true        -> zookeeper-exporter Go binary runs,
+  #                                   serving pull-mode metrics at :9095/metrics
+  #   metrics.enabled + pushMode  -> additionally pushes to Pushgateway via cron
+  #   metrics.fluentd.enabled     -> Fluentd runs independently for log forwarding
+  # The sidecar container spawns if EITHER metrics.enabled or fluentd.enabled is true.
   enabled: false
   image: radiantone/zk-exporter
   imageTag: latest
   zk_conn: localhost:2181
+  # NOTE: this securityContext is currently unused by the template — the sidecar
+  # uses the runtime default user (uid 1000 from Dockerfile USER fluent).
+  # Left for future wiring.
   securityContext:
     runAsUser: 0
   annotations: {}
-  pushMode: true
+  # pushMode: false by default. Only enable when a reachable Pushgateway exists.
+  # The exporter exits with code 1 after a 300s Pushgateway timeout, causing
+  # CrashLoopBackOff if pushMode=true but PUSHGATEWAY_URI is unreachable.
+  pushMode: false
   pushGateway: http://prometheus-pushgateway:9091
   pushMetricCron: "* * * * *"
   fluentd: