Huey

What Huey is

Huey is a lightweight, self-hosted task queue for Python applications. Its API supports tasks that run immediately, after a delay, at a chosen time, or on recurring crontab-like schedules. Tasks and results can be stored in Redis, Valkey, Redict, PostgreSQL, SQLite, the filesystem, or memory. Huey can retry failed tasks with configurable delays and backoff; other controls include prioritization, result storage, expiration, locking, rate limits, and timeouts. For composed work, it supports pipelines and chains, plus groups for fan-out and chords for map/reduce workflows. Consumers can run tasks using multiple processes, threads, or greenlets. Huey integrates with Django, including optional Django admin visibility and management, and documents Flask-Peewee admin integration and extensions for asyncio and task statistics. Immediate mode runs task functions synchronously for development and testing. Deployment guidance covers Docker, Docker Compose, Kubernetes, and PaaS environments, and the documentation includes a signed serializer recipe for untrusted environments. Huey is open source and costs nothing. It has no dependencies beyond the Python standard library, though Redis-like brokers and PostgreSQL need their respective client libraries.

Who it suits

Huey suits Python developers who need a self-hosted queue for scheduled or background tasks, retries, or composed workflows. Django users can also use its framework integration and optional admin tools.

Getting started

1. Open the Huey documentation website
2. Install Huey in a Python environment
3. Choose a storage backend such as Redis, PostgreSQL, SQLite, filesystem, or memory
4. Configure consumers with processes, threads, or greenlets
5. Deploy with Docker, Docker Compose, Kubernetes, or a PaaS environment

Verdict

Huey provides scheduling, retries, task controls, and workflow composition for Python projects, with framework integrations and multiple worker models. Its core has no dependencies beyond Python’s standard library, while some storage backends need additional client libraries.