Quickstart¶

1. Define tasks¶

from potatoq import Potatoq

app = Potatoq("proj")


@app.task
def add(x, y):
    return x + y


@app.task(bind=True, autoretry_for=(ConnectionError,), max_retries=5)
def fetch_avatar(self, user_id):
    ...

Potatoq("proj") finds its broker in this order:

1. broker= / app.conf.broker_url
2. $POTATOQ_BROKER_URL (or $CELERY_BROKER_URL)
3. your Django DATABASES["default"], if it's Postgres or SQLite
4. sqlite:///potatoq.sqlite3, with a warning. That's fine for development.

2. Start a worker¶

$ potatoq -A proj.tasks worker
[INFO/MainProcess] potatoq 26.1 worker potatoq@laptop ready: broker=sqlite:///potatoq.sqlite3
  results=sqlite:///potatoq.sqlite3 queues=default concurrency=8 (prefork) time_limit=1800s ...

The worker forks one process per available CPU. Each process runs one task at a time, so a slow task never holds others hostage. The same worker also runs your periodic tasks.

3. Send work¶

>>> from proj.tasks import add
>>> result = add.delay(2, 3)
>>> result.get(timeout=10)
5

On the database brokers (SQLite, Postgres) results are stored automatically, because writing them costs nothing there. On Redis and RabbitMQ you opt in, as in Celery:

app = Potatoq("proj", broker="redis://localhost", backend="redis://localhost")

4. Look around¶

$ potatoq -A proj.tasks status      # live workers and what they're running
$ potatoq -A proj.tasks queues      # waiting tasks per queue
$ potatoq -A proj.tasks dead list   # tasks that failed for good, and why

Where next¶