I have a Django v5.0.6 application (Python v3.12) in which I integrated Celery v5.4 to run asynchronous tasks, such as generating images via the OpenAI API, which takes about 2 minutes.

The entire setup is deployed on Heroku. Here is the content of my Procfile:

web: gunicorn mngrs_backend.wsgi --timeout 600
worker: celery -A mngrs_backend worker -l info --time-limit=7200

However, after 30 seconds, I get the following error in the logs:

I am looking to increase the delay to avoid a timeout.

Here are the attempts I made without success:

• Added --timeout 600 in the gunicorn command in the Procfile

• Added --time-limit=7200 in the celery command in the Procfile

• Added a timeout to the API request: requests.post(url, json=payload, headers=headers, timeout=600)

• Modified the configuration in celery.py

    app = Celery("mngrs_backend")
    app.conf.update(
        task_soft_time_limit=60 * 10,  
        task_time_limit=60 * 12,  
    )
  

• Added in the project settings:

    CELERY_BROKER_TRANSPORT_OPTIONS = {
        "visibility_timeout": 60 * 60 * 2,  # 2 hours
    }
  

Locally with Celery, I do not have any timeout issues. I contacted Heroku support, but they told me the problem is with my Celery configuration.

Does anyone have a solution?