Azure self-hosted runners: show logs on failure

.github/workflows/azure-login/action.yml

@@ -0,0 +1,43 @@
+name: Azure Login
+description: Logs into Azure using a service principal
+inputs:
+  credentials:
+    description: Your credentials in JSON format
+    required: true
+
+runs:
+  using: "composite"
+  steps:
+    - name: Process Azure credentials
+      uses: actions/github-script@v7
+      env:
+        AZURE_CREDENTIALS: ${{ inputs.credentials }}
+      with:
+        script: |
+          if (!process.env.AZURE_CREDENTIALS) {
+            core.setFailed('The AZURE_CREDENTIALS secret is required.')
+            process.exit(1)
+          }
+
+          const azureCredentials = JSON.parse(process.env.AZURE_CREDENTIALS)
+          const {clientId, clientSecret, tenantId, subscriptionId} = azureCredentials
+
+          core.setSecret(clientId)
+          core.exportVariable('AZURE_CLIENT_ID', clientId)
+
+          core.setSecret(clientSecret)
+          core.exportVariable('AZURE_CLIENT_SECRET', clientSecret)
+
+          core.setSecret(tenantId)
+          core.exportVariable('AZURE_TENANT_ID', tenantId)
+
+          core.setSecret(subscriptionId)
+          core.exportVariable('AZURE_SUBSCRIPTION_ID', subscriptionId)
+
+    - name: Azure Login
+      shell: bash
+      run: |
+        echo "Logging into Azure..."
+        az login --service-principal -u ${{ env.AZURE_CLIENT_ID }} -p ${{ env.AZURE_CLIENT_SECRET }} --tenant ${{ env.AZURE_TENANT_ID }}
+        echo "Setting subscription..."
+        az account set --subscription ${{ env.AZURE_SUBSCRIPTION_ID }} --output none

.github/workflows/cleanup-self-hosted-runners.yml

@@ -21,38 +21,16 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - uses: actions/checkout@v4
-    - name: Process Azure credentials
-      uses: actions/github-script@v7
-      env:
-        AZURE_CREDENTIALS: ${{ secrets.AZURE_CREDENTIALS }}
+    - name: Azure Login
+      uses: ./.github/workflows/azure-login
       with:
-        script: |
-          if (!process.env.AZURE_CREDENTIALS) {
-            core.setFailed('The AZURE_CREDENTIALS secret is required.')
-            process.exit(1)
-          }
-
-          const azureCredentials = JSON.parse(process.env.AZURE_CREDENTIALS)
-          const {clientId, clientSecret, tenantId, subscriptionId} = azureCredentials
-
-          core.setSecret(clientId)
-          core.exportVariable('AZURE_CLIENT_ID', clientId)
-
-          core.setSecret(clientSecret)
-          core.exportVariable('AZURE_CLIENT_SECRET', clientSecret)
-
-          core.setSecret(tenantId)
-          core.exportVariable('AZURE_TENANT_ID', tenantId)
+        credentials: ${{ secrets.AZURE_CREDENTIALS }}
 
-          core.setSecret(subscriptionId)
-          core.exportVariable('AZURE_SUBSCRIPTION_ID', subscriptionId)
     - name: Discover VMs to delete
       env:
         GH_APP_ID: ${{ secrets.GH_APP_ID }}
         GH_APP_PRIVATE_KEY: ${{ secrets.GH_APP_PRIVATE_KEY }}
       run: |
-        az login --service-principal -u ${{ env.AZURE_CLIENT_ID }} -p ${{ env.AZURE_CLIENT_SECRET }} --tenant ${{ env.AZURE_TENANT_ID }}
-        az account set --subscription ${{ env.AZURE_SUBSCRIPTION_ID }}
         active_vms=$(az vm list -g ${{ secrets.AZURE_RESOURCE_GROUP }} | jq -c '.[] | {name,timeCreated}')
         current_time=$(date +%s)
         one_hour_ago=$(($current_time - 3600))

.github/workflows/create-azure-self-hosted-runners.yml

@@ -144,11 +144,11 @@ jobs:
         )
 
         echo "AZURE_ARM_PARAMETERS=$AZURE_ARM_PARAMETERS" >> $GITHUB_ENV
-        
+
     - name: Azure Login
-      uses: azure/login@v2
+      uses: ./.github/workflows/azure-login
       with:
-        creds: ${{ secrets.AZURE_CREDENTIALS }}
+        credentials: ${{ secrets.AZURE_CREDENTIALS }}
 
     - uses: azure/arm-deploy@v2
       id: deploy-arm-template
@@ -158,16 +158,29 @@ jobs:
         template: ./azure-self-hosted-runners/azure-arm-template.json
         parameters: ./azure-self-hosted-runners/azure-arm-template-example-parameters.json ${{ env.AZURE_ARM_PARAMETERS }}
         scope: resourcegroup
+
+    - name: Show some more information on failure
+      if: failure()
+      run: |
+        echo "::group::VM status"
+        az vm get-instance-view --resource-group ${{ secrets.AZURE_RESOURCE_GROUP }} --name ${{ steps.generate-vm-name.outputs.vm_name }} --query "instanceView.statuses"
+        az vm get-instance-view --resource-group ${{ secrets.AZURE_RESOURCE_GROUP }} --name ${{ steps.generate-vm-name.outputs.vm_name }} --query "statuses"
+        echo "::endgroup::"
+
+        echo "::group::Deployment logs"
+        az group deployment show --resource-group ${{ secrets.AZURE_RESOURCE_GROUP }} --name deploy-${{ steps.generate-vm-name.outputs.vm_name }}
+        echo "::endgroup::"
+
+        echo "::group::Extension logs"
+        az vm extension show --resource-group ${{ secrets.AZURE_RESOURCE_GROUP }} --vm-name ${{ steps.generate-vm-name.outputs.vm_name }} --name CustomScriptExtension
+        echo "::endgroup::"
 
     - name: Show post-deployment script output
+      if: always()
       env:
         CUSTOM_SCRIPT_OUTPUT: ${{ steps.deploy-arm-template.outputs.customScriptInstanceView }}
       run: echo "$CUSTOM_SCRIPT_OUTPUT" | jq -r '.substatuses[0].message' 
 
     - name: Deallocate the VM for later use
       if: env.DEALLOCATE_IMMEDIATELY == 'true'
-      uses: azure/CLI@v2
-      with:
-        azcliversion: 2.64.0
-        inlineScript: |
-          az vm deallocate -n ${{ steps.generate-vm-name.outputs.vm_name }} -g ${{ secrets.AZURE_RESOURCE_GROUP }} --verbose
+      run: az vm deallocate -n ${{ steps.generate-vm-name.outputs.vm_name }} -g ${{ secrets.AZURE_RESOURCE_GROUP }} --verbose