One of the principles of microservices is that a process should only be responsible for doing a single piece of the workflow. A natural step while migrating from monoliths to microservices would be to define possible long-running tasks and split them into individual processes.
This is the concept behind task queues. Task queues handle the entire life cycle of managing tasks. Instead of implementing threading or multiprocessing on your own, with task queues, you delegate the task to be performed, which is then asynchronously handled by the task queue. The task may be performed on the same machine as the originating process but it may also run on a machine with dedicated requirements.
The tasks and their results are asynchronous, so there is no blocking in the main process. Examples of popular task queues in web development are Celery for Python, Sidekiq for Ruby, Kue for Node.js, and Machinery for Go. All of them can be used with Redis as a broker. Unfortunately,...
