arq VS Flask-RQ2

I need to process a lot of long running IO heavy jobs with background workers. I've been using ARQ for a while but decided to take a crack at writing my own distributed queue.

Lately I've been writing asynchronous python code and yes, the resource integration problem has come again. Because now from version 1.4 SQLAlchemy has become asynchronous a new boilerplate was created. Now, as a scheduler, I took a completely asynchronous Arq. Considering the specifics of the service, long I/O operations, it seems that the service turned out to be more optimal in asynchronous execution. I haven't measured the performance yet, but I think I'll write another post about it.

From experience i would not use threads for this or any background jobs. I would use Celery or Flask-RQ2 to be your workers, you will also probably end up using them to run other tasks as you encounter the need for other jobs. They both use Redis as a broker and job store and you can use Redis for other things like caching and so many other useful features. I kind of like RQ2 more then Celery because its a little simpler but Celery has a lot more to offer, more features. RQ2 has rq-dashboard for monitoring jobs and Celery has Flower.

However, you might need to access the app’s context like how you’d do so in the Flask-RQ2 extension by using ScriptInfo from flask.cli:

For the CLI part, i recommend going with Click. Its great and already part of flask and easy to use. Other CLI libraries work too but why have more libraries. For scheduling jobs you might want to look at Flask-APScheduler. i used it in one of my projects at work for a while but ended up needing something that could scale to many more workers so had to rip it out but i didnt have a problem with it otherwise. You might want to fork Flask-APScheduler so you can update some libraries because it hasnt been touched in 2 years now or you can just use APScheduler alone. Flask-RQ2 and Celery are pretty good too but the workers need to run in separate processes and need Redis. You could use Redislite i think and with celery you could use a database as your broker, i have done it but dont recommend it.