[Helm] Add ServiceMonitor to the helm chart

.github/workflows/main.yml

@@ -6,13 +6,15 @@ on:
   workflow_dispatch:
   push:
     branches: [main]
-    paths-ignore:
-      - "docs/**"
-      - "**.md"
+    # OPTIMIZE: We generate new images even on non src code changes, but this cost is okay for now
+    # paths-ignore:
+    #   - "docs/**"
+    #   - "**.md"
   pull_request:
-    paths-ignore:
-      - "docs/**"
-      - "**.md"
+    # paths-ignore:
+    #   - "docs/**"
+    #   - "**.md"
+
 
 env:
   # Even though we can test against multiple versions, this one is considered a target version.
@@ -151,3 +153,58 @@ jobs:
           cache-to: type=gha,mode=max
           build-args: |
             TARGET_GOLANG_VERSION=${{ env.TARGET_GOLANG_VERSION }}
+
+
+  # Run e2e tests on devnet if the PR has a label "e2e-devnet-test"
+  e2e-tests:
+    runs-on: ubuntu-latest
+    needs: build-images
+    if: contains(github.event.pull_request.labels.*.name, 'e2e-devnet-test')
+    env:
+      ARGO_HTTP1: true
+      ARGO_SECURE: true
+      ARGO_SERVER: ${{ vars.ARGO_SERVER }}
+    permissions:
+      contents: "read"
+      id-token: "write"
+
+    steps:
+      - id: "auth"
+        uses: "google-github-actions/auth@v1"
+        with:
+          credentials_json: "${{ secrets.ARGO_WORKFLOW_EXTERNAL }}"
+
+      - id: "get-credentials"
+        uses: "google-github-actions/get-gke-credentials@v1"
+        with:
+          cluster_name: "nodes-gcp-dev-us-east4-1"
+          location: "us-east4"
+
+      - id: "install-argo"
+        run: |
+          curl -sLO https://github.com/argoproj/argo-workflows/releases/download/v3.4.7/argo-linux-amd64.gz
+          gunzip argo-linux-amd64.gz
+          chmod +x argo-linux-amd64
+          mv ./argo-linux-amd64 /usr/local/bin/argo
+          argo version
+
+      - id: "wait-for-infra"
+        shell: bash
+        run: |
+          start_time=$(date +%s) # store current time
+          timeout=900 # 15 minute timeout in seconds
+
+          until argo template get dev-e2e-tests --namespace=devnet-issue-${{ github.event.pull_request.number }}; do
+              current_time=$(date +%s)
+              elapsed_time=$(( current_time - start_time ))
+              if (( elapsed_time > timeout )); then
+                  echo "Timeout of $timeout seconds reached. Exiting..."
+                  exit 1
+              fi
+              echo "Waiting for devnet-issue-${{ github.event.pull_request.number }} to be provisioned..."
+              sleep 5
+          done
+
+      - id: "run-e2e-tests"
+        run: |
+          argo submit --wait --log --namespace devnet-issue-${{ github.event.pull_request.number }} --from 'wftmpl/dev-e2e-tests' --parameter gitsha="${{ github.event.pull_request.head.sha }}"

charts/CHANGELOG.md

@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.0.0.5] - 2023-05-20
+
+- Added `ServiceMonitor` to the helm chart.
+
 ## [0.0.0.4] - 2023-05-12
 
 - Added `nodeType` parameter to the helm chart, which is now actor-agnostic.

charts/pocket/README.md

@@ -127,4 +127,5 @@ privateKeySecretKeyRef:
 | serviceAccount.annotations | object | `{}` | Annotations to add to the service account |
 | serviceAccount.create | bool | `true` | Specifies whether a service account should be created |
 | serviceAccount.name | string | `""` | The name of the service account to use. If not set and create is true, a name is generated using the fullname template |
+| serviceMonitor.enabled | bool | `false` | enable service monitor |
 | tolerations | list | `[]` |  |

charts/pocket/templates/service-monitor.yaml

@@ -0,0 +1,12 @@
+{{ if .Values.serviceMonitor.enabled }}
+apiVersion: monitoring.coreos.com/v1
+kind: ServiceMonitor
+metadata:
+    name: {{ include "pocket.fullname" . }}
+spec:
+  endpoints:
+  - port: metrics
+  selector:
+    matchLabels:
+      {{- include "pocket.selectorLabels" . | nindent 6 }}
+{{ end }}

charts/pocket/templates/statefulset.yaml

@@ -58,10 +58,12 @@ spec:
             - -config=/pocket/configs/config.json
             - -genesis=/pocket/configs/genesis.json
           ports:
-            - containerPort: 42069
+            - containerPort: {{ .Values.service.ports.consensus }}
               name: consensus
-            - containerPort: 50832
+            - containerPort: {{ .Values.service.ports.rpc }}
               name: rpc
+            - containerPort: {{ .Values.service.ports.metrics }}
+              name: metrics
           env:
           {{ if or .Values.privateKeySecretKeyRef.name .Values.privateKeySecretKeyRef.key }}
             - name: POCKET_PRIVATE_KEY

charts/pocket/values.yaml

@@ -170,6 +170,8 @@ service:
   annotations: {}
   # -- service type
   type: ClusterIP
+
+  # If adjusting values here, please make sure to also change the configuration of the pocket node (`config` section above)
   ports:
     # -- consensus port of the node
     consensus: 42069
@@ -178,6 +180,10 @@ service:
     # -- OpenTelemetry metrics port of the node
     metrics: 9000
 
+serviceMonitor:
+  # -- enable service monitor
+  enabled: false
+
 ingress:
   # -- enable ingress for RPC port
   enabled: false