Implement persistent background agent execution with scheduling and event-driven triggers

[Copilot]

Enables agents to run autonomously without UI dependency, supporting scheduled execution (cron/interval/date), webhook triggers, and event-driven workflows.

Implementation

Database Layer

• BackgroundAgent model: stores agent config, trigger settings, status (ACTIVE/PAUSED/STOPPED/ERROR)
• BackgroundAgentExecution model: tracks execution history with results and errors
• Support for 5 trigger types: CRON, INTERVAL, DATE, WEBHOOK, EVENT
• Migration creates tables with proper indexes and foreign keys

Service Layer

• BackgroundAgentService: APScheduler-based service managing agent lifecycle
• Auto-loads ACTIVE agents on startup
• Methods: start_agent(), stop_agent(), pause_agent(), resume_agent(), trigger_agent()
• Async execution via execute_flow_background() using existing FlowRunner
• Execution history tracking with status and error capture

API Endpoints

13 new endpoints under /api/v1/background_agents/:

• CRUD: POST /, GET /, GET /{id}, PATCH /{id}, DELETE /{id}
• Lifecycle: POST /{id}/start, POST /{id}/stop, POST /{id}/pause, POST /{id}/resume
• Operations: POST /{id}/trigger, GET /{id}/status, GET /{id}/executions
• User-scoped authorization, flow_id filtering support

Usage Example

# Create hourly data sync agent
agent = {
    "name": "Hourly Data Sync",
    "flow_id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
    "trigger_type": "INTERVAL",
    "trigger_config": {"hours": 1},
    "input_config": {"input_value": "Sync data"}
}

response = requests.post(
    "http://localhost:7860/api/v1/background_agents/",
    json=agent,
    headers={"Authorization": f"Bearer {token}"}
)

# Start the agent
agent_id = response.json()["id"]
requests.post(f"http://localhost:7860/api/v1/background_agents/{agent_id}/start")

Dependencies

• Added APScheduler>=3.10.4,<4.0.0 for scheduling infrastructure

Testing

• 21 service layer unit tests covering all operations and trigger types
• 17 API integration tests including auth/authz, CRUD, and lifecycle operations
• CodeQL security scan: 0 alerts

Documentation

• API Reference: endpoint specifications, request/response schemas, error handling
• User Guide: getting started, use cases, configuration reference, troubleshooting

Original prompt

Feature Request
Enhance Langflow Agents to support persistent background execution, ensuring they remain active even when the UI is closed or user is offline. Agents should operate as dedicated background processes/services, enabling scheduled, event-driven, and autonomous workflows.

Enhancement Details
Implement a background runner for agents that can be initiated independently of the UI session.
Support scheduling capabilities (cron-like, time-based triggers).
Allow agents to respond to external events, webhooks, or API calls while running background tasks.
Provide controls for monitoring, starting, stopping, or restarting agent processes from the Langflow UI or API.
Ensure robust logging and error reporting for background agent activities.
Compatibility considerations: make background agent mode configurable per agent, with system resource checks and cross-platform support.
Motivation
This enhancement enables always-on and event-driven automation, supporting use cases such as:

Workflow orchestration (multi-step, autonomous tasks)
Continuous monitoring and alerting
Integration with third-party services or schedules
Reduces dependency on manual UI actions for agent activity
Your Contribution
I can assist with:

Use case validation: identifying scenarios that benefit from background agents
Testing: verifying background execution reliability on different platforms
Documentation: contributing guides and design suggestions for implementation
Collaboration: discussing system compatibility, API interface, and operational requirements
Next Steps
Gather feedback on specific background agent use cases and technical constraints
Prototype background agent runner and scheduling functionality
Document workflow and API integration options
Collaborate with community to ensure robust and user-friendly implementation

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.