When the task complexity rises further, splitting out the message queue and workers into separate processes makes sense again. That way, you can easily scale wo

When the task complexity rises further, splitting out the message queue and workers into separate processes makes sense again. That way, you can easily scale worker capacity up and down and you can restart your web server processes without having to worry about background job persistence since that is handled by the message queue.