[Python] Celery 5.5.2 randomly raises RecursionError when the AsyncResult.status property is...

I have this simple line of code:

task_id = payload.task_id
task_result = AsyncResult(task_id)
if task_result.status == "SUCCESS":
print("I am happy")


Celery configuration is pretty basic. I just set the broker and the backend to use the same RabbitMQ instance; any other setting is left to default.

I wasn't able to reproduce. I could only produce two consecutive times using a random id (123). A few moments later, it normally worked, returning PENDING

Looking at the stack trace, I saw a loop around those instructions

self.connection.drain_events(timeout=timeout)
while not self.blocking_read(timeout):
return self.on_inbound_frame(frame)
callback(channel, method_sig, buf, None)
return self.channels[channel_id].dispatch_method(
listener(*args)
self._do_revive()
self.open()
return self.send_method(
return self.wait(wait, returns_tuple=returns_tuple)
self.connection.drain_events(timeout=timeout)

Continue reading...