The problem is this: when we push tasks to the queue, there are inevitably many kinds of tasks to push, such as e-mail, short messages, etc. when the backstage work process takes tasks, we are often locked in thinking, thinking that I push tasks according to categories, then whether to take tasks according to categories, then it is possible to have multiple task processing processes at this time. in fact, the only difference between them is the parameters of the task categories. Isn’t this a bad design?
I still suggest to use different queues to process different information. Although the parameters can be used separately, it is not conducive to isolation and expansion in subsequent large concurrent events. For example, a certain service only focuses on short message information, so it is sufficient to monitor the short message queue.