Lifespan Tasks

Added in v0.5.2

Lifespan tasks are coroutines that run when the backend server is running. They are useful for setting up the initial global state of the app, running periodic tasks, and cleaning up resources when the server is shut down.

Lifespan tasks are defined as async coroutines or async contextmanagers. To avoid blocking the event thread, never use time.sleep or perform non-async I/O within a lifespan task.

In dev mode, lifespan tasks will stop and restart when a hot-reload occurs.

Any async coroutine can be used as a lifespan task. It will be started when the backend comes up and will run until it returns or is cancelled due to server shutdown. Long-running tasks should catch asyncio.CancelledError to perform any necessary clean up.

To register a lifespan task, use app.register_lifespan_task(coro_func, **kwargs). Any keyword arguments specified during registration will be passed to the task.

If the task accepts the special argument, app, it will be an instance of the FastAPI object associated with the app.

Lifespan tasks can also be defined as async contextmanagers. This is useful for setting up and tearing down resources and behaves similarly to the ASGI lifespan protocol.

Code up to the first yield will run when the backend comes up. As the backend is shutting down, the code after the yield will run to clean up.

Here is an example borrowed from the FastAPI docs and modified to work with this interface.