Backend async tasks (Celery)

When to use

• You are adding or changing a Celery task in backend/src/backend/apps/**/tasks.py.
• You need to update the Celery Beat schedule in backend/src/backend/celery_app.py.
• You are adding DB access inside a Celery worker process.

Steps

1. Define tasks only in a tasks.py file under the relevant app module.
2. Decorate tasks with @app.task imported from backend.celery_app.
3. Return a typed result (int, None, etc.). Keep task-specific parameters

local unless they must be environment-configured.

1. For DB access:

- Use asyncio.run() to bridge async work. - Always acquire a session via gettasksession() from backend.celery_app.

1. For periodic jobs, update app.conf.beat_schedule in

backend/src/backend/celery_app.py and use celery.schedules.crontab.

Constraints and guardrails

• Never use the global alchemy_config engine inside a task; it is not fork-safe

for Celery workers.

• Always use gettasksession() to manage DB connections safely across worker

processes.

References

• backend/src/backend/apps/**/tasks.py
• backend/src/backend/celery_app.py
• backend/tests/apps/users/test_tasks.py