>>anonu+(OP)
This is fantastic for relatively low volume queues with intensive work to be done by a small number of workers. For these types of use cases, I'll take this approach over RabbitMQ or Kafka all day long.
But once you get even just up to say, 40 messages per second with 100 worker processes, you're now up to 4000 updates per second just to see which worker got to claim which job, and up from there becomes untenable.