Hi, when I run a python script like this:

if __name__ == "__main__":
    upload_deploy = upload_flow.to_deployment("upload-file-deploy", work_pool_name="my-pool")
    transform_deploy = transform_flow.to_deployment("transform-deploy", interval=30, work_pool_name="my-pool")

    serve(upload_deploy, transform_deploy)

Is this process considered a worker? I'm running everything locally for now. It seems to be able to run the scheduled tasks and it also responds to "run_deployment" calls from other processes. I have not been able to wrap my head around how "prefect worker start" fits into this.

As you noted above, it’s my understanding with serve you are using the local resources of the machine where you are running that command. I can only speak to using prefect cloud, but the workers come into play at that point, or when you deploy elsewhere.

My goal is to get to a self-hosted solution, so want to understand whether I'd need a dedicated worker container or if I could simply spin up a "serve" container on a single node.

Hey I would checkout the docs if you haven't already, particularly these: • https://docs.prefect.io/v3/deploy/run-flows-in-local-processes • https://docs.prefect.io/v3/deploy/infrastructure-concepts/work-pools Just to hopefully give a little more context: Serving a flow is not the same as running a worker. You absolutely can simply serve() a flow on a single node. As you mentioned above you get scheduling and the ability to run on demand (run deployment). However these run are spawned into individual processes on that same machine. A worker performs a similar function but it can launch runs into the configured infrastructure such as ECS, Cloud Run, etc. where you may need more seperation and control over the runtime environment. Hope that helps!

Thanks Jake, that helps for sure!