Got it — you’ve got a deployment configured with a specific work queue. What would you like to do next: start a worker that polls that queue, or troubleshoot runs that aren’t being picked up? If you can share your deployment name, work pool, work queue, and the exact worker start command you’re using, I can be precise. Quick reference for Prefect 3.x: - A deployment can target a specific work queue via

work_queue_name

. - A worker can poll one or more specific queues with `-q/--work-queue`; if you omit

-q

, it will poll all queues in the pool. Start a worker for a specific pool/queue

prefect worker start --pool my-pool -q my-queue

(repeat

-q

to poll multiple queues) Create a deployment (Python) that targets a queue

from prefect import flow

@flow
def my_flow():
    ...

if __name__ == "__main__":
    my_flow.deploy(
        name="my-deployment",
        work_pool_name="my-pool",
        work_queue_name="my-queue",
    )

Run the deployment

prefect deployment run "my_flow/my-deployment" --watch

Inspect the deployment (to confirm its queue)

prefect deployment inspect "my_flow/my-deployment"

Notes and differences from 2.x - Agents were replaced by workers; use

prefect worker start

instead of

prefect agent start

. -

Deployment.build_from_flow()

and

prefect deployment build

were removed; use

.deploy()

in code or

prefect deploy

. - Workers can be restricted to specific queues via repeated

-q

flags. Docs for more detail: - Work pools and queues in Prefect 3.x - Deploy via Python (`flow.deploy` with `work_queue_name`) - Worker CLI reference If you’re seeing runs stay in Pending, check: - The worker is running on the correct pool and queue (

--pool

and

-q

match). - The queue isn’t paused. - The worker type matches the work pool type (e.g., Docker/Kubernetes). - Any queue filters (tags/deployment IDs) aren’t excluding your deployment. Share your deployment snippet and the worker command you’re using and I’ll help you wire it up.

Nothing helps, the work_queue_id is still None

This was a defect, now seems to be fixed; I am waiting for the release