v2.0.0 - Unified Monitoring Endpoint

Release Notes

v2.0.0 - Unified Monitoring Endpoint (2025-01-21)

🚀 Major Changes

Unified SSE Monitoring

• Migrated to unified SSE monitoring endpoint (/api/v1/monitor/stream/unified)
• Single efficient stream for both tests and workflows
• Reduced server load and improved performance
• Better real-time event delivery

Enhanced Failure Detection

• Workflows that fail now correctly trigger workflow_failed events
• Tests that fail now correctly trigger test_failed_with_data events
• Proper status reporting in GitHub Actions outputs
• Failed workflows/tests no longer appear as successful

Improved Error Handling

• Better SSE connection error messages
• Graceful handling of connection drops
• Non-JSON error event handling
• More descriptive error outputs

🔧 Technical Improvements

Backend Integration

• Updated to use GlobalStreamingContext on backend
• Redis pub/sub for event-driven updates (no polling)
• Cached initial state for faster connection
• Race condition fix: subscribes before sending initial state

Event Handling

• connection_ready - Connection established
• initial_state - Current running tests and workflows
• workflow_started, workflow_updated - Real-time progress
• workflow_completed, workflow_failed, workflow_cancelled - Terminal
  states
• test_started, test_updated - Test progress
• test_completed_with_data, test_failed_with_data - Test results
• heartbeat - Connection keep-alive

⚠️ Breaking Changes

Backend Requirements

• Requires backend with unified monitoring service
• Old endpoints no longer supported:
  • ❌ /api/v1/tests/monitor/stream
  • ❌ /api/v1/workflows/monitor/stream
  • ✅ /api/v1/monitor/stream/unified (new)

Migration Guide No action needed if using latest backend. The action
automatically uses the new endpoint.

🐛 Bug Fixes

• Fixed workflow failures being incorrectly reported as successful
• Fixed test failures being incorrectly reported as completed
• Improved SSE connection stability
• Better handling of edge cases in initial state parsing
• Fixed race condition in workflow monitoring initialization

📊 Outputs

All existing outputs remain the same:

• task_id - Unique task identifier
• success - true/false completion status
• status - Final status (completed/failed/cancelled)
• total_tests - Total tests in workflow
• completed_tests - Tests completed
• passed_tests - Tests that passed
• failed_tests - Tests that failed
• error_message - Error details on failure
• report_link - Link to test/workflow report

🎯 Usage

- name: Run Revyl Workflow
  uses: RevylAI/revyl-gh-action/run-test@v2.0.0
  with:
    workflow-id: 'your-workflow-id'
    timeout: '3600'
  env:
    REVYL_API_KEY: ${{ secrets.REVYL_API_KEY }}

Or use @main for latest:

uses: RevylAI/revyl-gh-action/run-test@main

📝 Related Changes

• Backend: Unified monitoring service implementation
• Frontend: GlobalStreamingContext integration
• Schemas: OrgWorkflowMonitorItem and OrgTestMonitorItem types

---

Previous Releases

v1.x - Legacy Monitoring

• Separate SSE endpoints for tests and workflows
• Polling-based status updates
• Basic error handling

v1.1.1

What's Changed

• updating README by @anamhira47 in #6

Full Changelog: v1.1...v1.1.1

v1.1

What's Changed

• Releases/v1 by @anamhira47 in #1

Full Changelog: https://github.com/CogniSim/cognisim-gh-action/commits/1.0

What's Changed

• removing code owners by @anamhira47 in #2
• updating action.yml by @anamhira47 in #3

Full Changelog: 1.0...v1.1