Usage

from redis import Redis
from rediscron import RedisCronScheduler


def rebuild_metric(metric_id: str) -> None:
    ...


connection = Redis.from_url("redis://localhost:6379/0")
scheduler = RedisCronScheduler(connection)
scheduler.register(
    rebuild_metric,
    queue_name="metrics",
    args=("pm25",),
    interval=300,
)
scheduler.start()

Redis keys

Keys used by this package:

• rq:cron_job:{id} stores one cron job hash.

• rq:cron_jobs indexes job ids by next enqueue timestamp.

• rq:cron_jobs:lock allows one live scheduler process at a time.

• rq:cron_jobs:last_update records runtime changes.

• rq:cron_jobs:events publishes create, update, and delete events.

Development

To install pre-commit, do the following:

$ uv tool install pre-commit --with pre-commit-uv

To enable pre-commit git hooks:

$ pre-commit install

Why?

The original cron implementation in RQ is holding all the cron jobs locally as a list inside the instance of the scheduler. Because of this design choice, as soon as the scheduler is started, the list of scheduled jobs cannot be easily modified.

In some cases, we want to be able to add/edit/delete scheduled jobs on the fly and for the scheduler to acknowledge those events accordingly. This is what lead the design choices for this package.

Inspired by redisbeat, a celery scheduler using redis to store metadata. See Hello RedBeat: A New Celery Beat Scheduler for mor info.