django-orchestra/orchestra/contrib/tasks/README.md

7 lines
731 B
Markdown
Raw Normal View History

This is a wrapper around djcelery and celery `@task` and `@periodic_task` decorators. It provides transparent support for switching between executing a task on a plain Python thread or
2015-05-07 20:04:28 +00:00
the traditional way of pushing the task on a queue (rabbitmq) and wait for a Celery worker to run it.
2015-10-15 22:31:54 +00:00
A queueless threaded execution has the advantage of 0 moving parts instead of the alternative rabbitmq and celery workers. Less dependencies, less memory footprint, less points of failure, no process keeping, no independent code reloading for the workers.
If your application needs to run thousands or milions of tasks a day, use celery as your backend, if tens or hundreds, then probably the default thread backend will be your best choice.