Playbook Execution

The subscribe_pending_playbook background task is responsible for discovering and executing pending Playbooks.

Discovery Mechanism

PlaybookLoader (Lib/playbookloader.py) scans the PLAYBOOKS/ directory at startup, dynamically imports each .py file, reads the NAME and DESC properties of the Playbook class, and caches them in Xcache.

Execution Flow

API/MCP creates Playbook record (job_status=PENDING)
  └── subscribe_pending_playbook polls and discovers
        ├── Looks up Playbook configuration in Xcache
        ├── importlib dynamically loads the Playbook class
        ├── Sets job_status=RUNNING, generates job_id
        └── thread_module_manager.start_task() dispatches thread
              └── playbook.execute() → playbook.run()
                    └── On completion, callback on_playbook_task_finished()

Status Fallback

The on_playbook_task_finished callback ensures that every Playbook reaches a terminal state:

• If the business code has already set a terminal state via update_playbook_status(), the callback skips
• If the business code has not set one, the callback automatically writes SUCCESS or FAILED

Key APIs

The Playbook base class (Lib/baseplaybook.py) provides:

API	Description
self.param_source_row_id	The Case row_id that triggered the Playbook
self.param_user_input	Additional user input
self.update_playbook_status(status, remark)	Update task status
self.send_notice(title, body)	Send a notification to the user