Short answer: it should work, but there was a bug. In Prefect 3.x, if you omit

work_queue_name

when calling

create_flow_run_from_deployment

, the server is supposed to use the deployment’s configured work queue by default. If the deployment is not associated with a work queue (only a work pool), or you’re on a version with the bug below, the run won’t be routed to a queue. Known issue: - A bug in 3.4.11 caused

create_flow_run_from_deployment

to mis-handle deployments without an attached work queue and could raise an AttributeError; the fix added proper checks and clearer errors. See: - Issue: #18707 - Fix: #18807 What I’d suggest: - Confirm your Prefect version and upgrade to the latest 3.4.x if you’re on 3.4.11 or earlier. - Ensure your deployment actually has a specific work queue set in its work pool (not just the pool). If no queue is set on the deployment, omitting

work_queue_name

won’t target a queue. - As a workaround, pass

work_queue_name

explicitly to

create_flow_run_from_deployment

. A couple of clarifying questions to help pinpoint this for you: - Which Prefect version are you on? - Does the deployment show a specific “Work pool / Work queue” in the UI? - Are you using durable work queues (e.g., Redis) and, if so, how are you checking whether the message hit the queue?