Celery¶

Thelogfire.instrument_celery() method will create a span for every taskexecuted by your Celery workers.

The integration also supports theCelery beat.

Installation¶

Installlogfire with thecelery extra:

pipinstall'logfire[celery]'

uvadd'logfire[celery]'

Celery Worker¶

For our example, we'll useredis. You can run it with Docker:

dockerrun--rm-d-p6379:6379redis

Below we have a minimal example using Celery. You can run it withcelery -A tasks worker --loglevel=info:

importlogfirefromceleryimportCeleryfromcelery.signalsimportworker_init@worker_init.connect()# (1)!definit_worker(*args,**kwargs):logfire.configure(service_name="worker")# (2)!logfire.instrument_celery()app=Celery("tasks",broker="redis://localhost:6379/0")# (3)!@app.taskdefadd(x:int,y:int):returnx+yadd.delay(42,50)# (4)!

1. Celery implements different signals that you can use to run code at specific points in the application lifecycle. You can see more about the Celery signalshere.
2. Use aservice_name to identify the service that is sending the spans.
3. Installredis withpip install redis.
4. Trigger the task synchronously. On your application, you probably want to useapp.send_task("tasks.add", args=[42, 50]). Which will send the task to the broker and return immediately.

Celery Beat¶

As said before, it's also possible that you have periodic tasks scheduled withCelery beat.

Let's add the beat to the previous example:

importlogfirefromceleryimportCeleryfromcelery.signalsimportworker_init,beat_init@worker_init.connect()definit_worker(*args,**kwargs):logfire.configure(service_name="worker")logfire.instrument_celery()@beat_init.connect()# (1)!definit_beat(*args,**kwargs):logfire.configure(service_name="beat")# (2)!logfire.instrument_celery()app=Celery("tasks",broker="redis://localhost:6379/0")app.conf.beat_schedule={# (3)!"add-every-30-seconds":{"task":"tasks.add","schedule":30.0,"args":(16,16),},}@app.taskdefadd(x:int,y:int):returnx+y

1. Thebeat_init signal is emitted when the beat process starts.
2. Use a differentservice_name to identify the beat process.
3. Add a task to the beat schedule. See more about the beat schedulehere.

The code above will schedule theadd task to run every 30 seconds with the arguments16 and16.

To run the beat, you can use the following command:

celery-Atasksbeat--loglevel=info

The keyword arguments oflogfire.instrument_celery() are passed to theCeleryInstrumentor().instrument() method.