What is everyone using for scheduled background jobs and queued tasks? (not celery)

[epicwhale]

I am wondering what's a modern task queue that's lightweight and also couples well with the rest of this full stack template? For both scheduled jobs and queued tasks that arise from the backend.

• I have explored arq so far but I find it rather unmaintained recently, lacking strong observability and UI-dashboard visibility into jobs running / scheduled - even though I appreciate its simplicity.

• celery requires hacking around for async compatibility - if the code i want to run in worker tasks is async heavy

• i was considering prefect + sqlite as a third alternative, but comes with a learning curve

what are others here doing?

[bysiber]

Had the same itch, ended up building one. flashq uses SQLite as the backend, zero dependencies, no broker to set up.

from flashq import FlashQ

app = FlashQ()

@app.task
def send_report(user_id):
    ...

# queue it
result = send_report.delay(user_id=42)

# or schedule it
result = send_report.delay(user_id=42, countdown=3600)  # runs in 1 hour

Supports retries with exponential backoff, task chaining (Chain/Group/Chord), and middleware hooks. Tasks persist in SQLite so they survive restarts.

No dashboard UI yet, but since everything sits in a .db file you can inspect task states directly. Main limitation: single-node only, no distributed workers. For a typical FastAPI deployment that shouldn't matter though.

pip install flashq | repo: https://github.com/bysiber/flashq

[epicwhale]

I wonder if someone could build a robust sqlite adapter for taskiq instead? As it feels like the most promising and stable / modern / proven celery and arq alternative with true async support.

The reason I say this as there are always new background tasks systems for redis that spin up and then remain unmaintained a few months down the road..