Here is my system loaded up with work:
It seems Broadway is favoring several workers and giving them lots of messages to process (when I have max_demand: 1). Others are getting no messages, so turning up the number of workers has no effect on throughput. CPU & Network never hit maximum, the system just processes slowly.
Does Broadway use a round-robin message distribution strategy? Can I change it? How would I ensure each worker only ever has 1 message to work on at a time?
