Frequently asked questions

FastWorker is an open-source (MIT) brokerless task queue for Python. It runs distributed background jobs using a control-plane process and direct NNG messaging instead of an external broker like Redis or RabbitMQ.

Correct. FastWorker has no external broker and no separate result backend. Coordination happens over NNG between a control plane and one or more subworkers — typically 2–3 Python processes in total.

It covers the same core workflow — define tasks, enqueue them, run workers — but with a different API (an async client and a control-plane/subworker model). Moving from Celery is a guided migration rather than a byte-for-byte swap. See the Celery migration guide.

FastWorker ships an async-native Client. You call await client.delay("task", ...) directly inside async FastAPI request handlers — no synchronous wrapper and no thread-pool bridge.

It targets Python-only stacks running roughly 1,000–10,000 tasks per minute across 1–50 workers. For extreme fan-out, very large fleets, or polyglot workers, a broker-based system may fit better.

Four priority levels are built in — critical, high, normal, and low. Priority is a first-class argument on a task, and the control plane dispatches each task to the least-loaded subworker at the highest available priority.

Yes. A real-time management dashboard ships with the control plane and auto-starts at http://127.0.0.1:8080. It shows live workers, queue depth by priority, task history, and result-cache stats — no Flower or rq-dashboard required.

Yes. FastWorker offers optional OpenTelemetry instrumentation that emits spans for submission, dispatch, and execution to any OTLP collector such as Jaeger or Honeycomb.

FastWorker requires Python 3.12 or newer. It is currently in alpha (v0.1.1) and MIT licensed. Install it with pip install fastworker or uv add fastworker.

The control plane keeps a configurable result cache (default 10,000 entries with a one-hour TTL). Expired entries are swept every minute. Both size and TTL are tunable via --result-cache-size and --result-cache-ttl.

FastWorker is a pure-Python package and targets Linux, macOS, and Windows wherever Python 3.12+ and NNG are available. Linux and macOS are the primary tested platforms.

FastWorker is currently alpha software (v0.1.1). The core control-plane, dispatch, and dashboard work today, but APIs may change before 1.0. Evaluate it against your reliability needs and pin your version.