Update triage agent workflow to start orchestration

.github/workflows/triage-agent.yml

@@ -49,8 +49,8 @@ jobs:
             const labelNames = (issue.labels || []).map(label => label.name);
             core.setOutput('labels', JSON.stringify(labelNames));
 
-      - name: Call Azure Function
-        id: call_azure_function
+      - name: Start Triage Orchestration
+        id: start_orchestration
         env:
           PAYLOAD: >-
             {
@@ -67,13 +67,14 @@ jobs:
             }
 
         run: |
-          # Make the HTTP request with improved error handling and timeouts
-          echo "Making request to triage agent..."
+          # Start the durable function orchestration
+          echo "Starting triage orchestration..."
+          echo "Payload size: $(echo "$PAYLOAD" | wc -c) bytes"
 
-          # Add timeout handling and better error detection
+          # Make the initial request to start orchestration
           set +e  # Don't exit on curl failure
-          response=$(timeout ${{ vars.TRIAGE_AGENT_TIMEOUT }} curl \
-              --max-time 0 \
+          response=$(curl \
+              --max-time 60 \
               --connect-timeout 30 \
               --fail-with-body \
               --silent \
@@ -82,19 +83,16 @@ jobs:
               --header "Content-Type: application/json" \
               --request POST \
               --data "$PAYLOAD" \
-              ${{ secrets.TRIAGE_FUNCTION_LINK }} 2>&1)
+              ${{ secrets.TRIAGE_START_ENDPOINT }} 2>&1)
 
           curl_exit_code=$?
           set -e  # Re-enable exit on error
 
-          echo "Curl exit code: $curl_exit_code"
+          echo "Start orchestration curl exit code: $curl_exit_code"
 
-          # Check if curl command timed out or failed
-          if [ $curl_exit_code -eq 124 ]; then
-            echo "❌ Request timed out after 650 seconds"
-            exit 1
-          elif [ $curl_exit_code -ne 0 ]; then
-            echo "❌ Curl command failed with exit code: $curl_exit_code"
+          # Check if curl command failed
+          if [ $curl_exit_code -ne 0 ]; then
+            echo "❌ Failed to start orchestration with exit code: $curl_exit_code"
             echo "Response: $response"
             exit 1
           fi
@@ -112,11 +110,115 @@ jobs:
             exit 1
           fi
 
-          # Check if the request was successful
+          # Check if the orchestration started successfully
           if [ "$http_code" -ge 200 ] && [ "$http_code" -lt 300 ]; then
-            echo "✅ Azure Function call succeeded"
+            echo "✅ Triage orchestration started successfully"
+            echo "Response: $response_body"
+
+            # Extract instance ID from response (assuming JSON response with instanceId field)
+            instance_id=$(echo "$response_body" | grep -o '"instanceId":"[^"]*"' | cut -d'"' -f4)
+            if [ -z "$instance_id" ]; then
+              echo "❌ Failed to extract instance ID from response"
+              echo "Response body: $response_body"
+              exit 1
+            fi
+
+            echo "Instance ID: $instance_id"
+            echo "instance_id=$instance_id" >> $GITHUB_OUTPUT
           else
-            echo "❌ Azure Function call failed with status code: $http_code"
+            echo "❌ Failed to start orchestration with status code: $http_code"
             echo "Response: $response_body"
             exit 1
           fi
+
+      - name: Poll Orchestration Status
+        id: poll_status
+        run: |
+          # Poll the orchestration status until completion
+          instance_id="${{ steps.start_orchestration.outputs.instance_id }}"
+          echo "Polling status for instance: $instance_id"
+
+          max_attempts=120  # Maximum number of polling attempts (10 minutes with 5-second intervals)
+          attempt=0
+
+          while [ $attempt -lt $max_attempts ]; do
+            attempt=$((attempt + 1))
+            echo "Polling attempt $attempt/$max_attempts..."
+
+            # Make status check request
+            set +e  # Don't exit on curl failure
+            # Safely replace the {instanceId} placeholder using bash parameter expansion
+            status_url="${{ secrets.TRIAGE_STATUS_ENDPOINT }}"
+            status_url="${status_url//\{instanceId\}/$instance_id}"
+
+            response=$(curl \
+                --max-time 30 \
+                --connect-timeout 15 \
+                --fail-with-body \
+                --silent \
+                --show-error \
+                --write-out "HTTPSTATUS:%{http_code}" \
+                --request GET \
+                "$status_url" 2>&1)
+
+            curl_exit_code=$?
+            set -e  # Re-enable exit on error
+
+            if [ $curl_exit_code -ne 0 ]; then
+              echo "❌ Status check failed with exit code: $curl_exit_code"
+              echo "Response: $response"
+              sleep 5
+              continue
+            fi
+
+            # Extract HTTP status code and response body
+            http_code=$(echo "$response" | grep -o "HTTPSTATUS:[0-9]*" | cut -d: -f2)
+            response_body=$(echo "$response" | sed 's/HTTPSTATUS:[0-9]*$//')
+
+            echo "Status check HTTP code: $http_code"
+
+            if [ "$http_code" -ge 200 ] && [ "$http_code" -lt 300 ]; then
+              echo "Status response: $response_body"
+
+              # Check if orchestration is complete (using runtimeStatus field from actual response schema)
+              runtime_status=$(echo "$response_body" | grep -o '"runtimeStatus":"[^"]*"' | cut -d'"' -f4)
+
+              if [ "$runtime_status" = "Completed" ] || [ "$runtime_status" = "completed" ]; then
+                echo "✅ Triage orchestration completed successfully"
+                # Extract and display output if available
+                output=$(echo "$response_body" | grep -o '"output":[^,}]*' | cut -d':' -f2-)
+                if [ "$output" != "null" ] && [ -n "$output" ]; then
+                  echo "Orchestration output: $output"
+                fi
+                exit 0
+              elif [ "$runtime_status" = "Failed" ] || [ "$runtime_status" = "failed" ]; then
+                echo "❌ Triage orchestration failed"
+                echo "Final response: $response_body"
+                # Extract and display output for error details if available
+                output=$(echo "$response_body" | grep -o '"output":[^,}]*' | cut -d':' -f2-)
+                if [ "$output" != "null" ] && [ -n "$output" ]; then
+                  echo "Error details: $output"
+                fi
+                exit 1
+              elif [ "$runtime_status" = "Running" ] || [ "$runtime_status" = "running" ] || [ "$runtime_status" = "Pending" ] || [ "$runtime_status" = "pending" ]; then
+                echo "Orchestration status: $runtime_status, waiting..."
+                # Display last updated time for better monitoring
+                last_updated=$(echo "$response_body" | grep -o '"lastUpdatedAt":"[^"]*"' | cut -d'"' -f4)
+                if [ -n "$last_updated" ]; then
+                  echo "Last updated: $last_updated"
+                fi
+                sleep 5
+              else
+                echo "Unknown runtime status: $runtime_status, continuing to poll..."
+                echo "Full response: $response_body"
+                sleep 5
+              fi
+            else
+              echo "❌ Status check returned HTTP $http_code"
+              echo "Response: $response_body"
+              sleep 5
+            fi
+          done
+
+          echo "❌ Orchestration polling timed out after $max_attempts attempts"
+          exit 1