This project is a complete worker and client implementation for the Faktory job server. You can use it to either consume jobs from Faktory or push jobs to the Faktory server to be processed.

Requires Python 3.7+.

❌ 0.5.0
✅ 0.6.0+
✅ 0.7.0+
✅ 0.8.0+
✅ 1.0.0+

• Creating a worker to run jobs from Faktory
• Concurrency (with multiple processes or threads with the use_threads=True option)
• Pushing work to Faktory from Python (with retries, custom metadata and scheduled support)
• Pushing exception / errors from Python back up to Faktory
• Sends worker status back to Faktory
• Supports quiet and teminate from the Faktory web UI
• Password authentication
• TLS support
• Graceful worker shutdown (ctrl-c will allow 15s for pending jobs to finish)

• Documentation (in progress, help would be appreciated)
• Tests (in progress, help would be appreciated)
• Django integration (./manage.py runworker and app/tasks.py support)

pip install faktory

There is a client context manager that you can use like this:

import faktory

with faktory.connection() as client:
    client.queue('test', args=(1, 2))
    client.queue('test', args=(4, 5), queue='other')

test doesn't need to be implemented by the Python worker, it can be any of the available worker implementations.

Sample worker:

from faktory import Worker

def your_function(x, y):
    return x + y

w = Worker(queues=['default'], concurrency=1)
w.register('test', your_function)

w.run()  # runs until control-c or worker shutdown from Faktory web UI

faktory_worker_python uses this format for the Faktory URL:

tcp://:password@localhost:7419

or with TLS:

tcp+tls://:password@localhost:7419

If the environment variable FAKTORY_URL is set, that is used. Otherwise you can pass the server URL in to the Worker or Client constructor, like this:

w = Worker(faktory="tcp://localhost:7419")

The worker users Python's built in logging module, which you can enable like this before calling .run():

import logging
logging.basicConfig(level=logging.DEBUG)