<@ULVA73B9P> I added a temp worker to pool due to ...
# ask-marvin
v
@Marvin I added a temp worker to pool due to high demand. how can I check if this new worker is actually help or if any tasks are going into this new worker? both workers are on two different systems.
m
thought for 303.9 seconds
Great question — and nice move adding capacity. Quick note: in Prefect 3.x, workers pull entire flow runs (not individual tasks). Here’s how to tell if your temporary worker is actually picking up runs. UI - Go to your pool: Work Pools → select your pool. - Workers tab: you should see both workers, their status, and last heartbeat/activity. If your temp worker is active, its heartbeat will update every few seconds. - Runs tab: filter by this pool to see throughput/backlog trends. If the backlog is dropping faster since you added the worker, it’s helping. - On a specific flow run’s page, check the logs. You’ll typically see lines indicating when a worker picked up/submitted the run. CLI - Inspect the pool to confirm it sees active workers and queues:
Copy code
prefect work-pool inspect "<your-pool>"
This helps confirm the worker is attached to the right pool and queues. - For any recent flow run, view logs and look for lines indicating submission/pickup by a worker:
Copy code
prefect flow-run logs <flow_run_id>
Look for lines like “Picked up flow run …” or “Submitted flow run …”; those usually include the worker name. On the worker host(s) - Start workers with a distinct name so you can identify them:
Copy code
prefect worker start --pool "<your-pool>" --name "temp-worker-xyz" [--work-queue "<queue-name>"]
- Watch the worker’s stdout. You should see logs like “Polling for scheduled flow runs…”, and when it’s helping you’ll see messages about picking up/submitting flow runs. If your temp worker appears idle - Make sure it’s pointed at the same pool and the correct work queue(s). If your deployments target specific queues, the worker must include those queue names via
--work-queue
. - Ensure the pool isn’t paused, and there isn’t a pool or queue-level concurrency limit preventing additional runs. - Confirm your deployments are assigned to this pool (and queues). - Increase worker log verbosity temporarily to see polling and claim attempts:
Copy code
PREFECT_LOGGING_LEVEL=DEBUG prefect worker start --pool "<your-pool>" --name "temp-worker-xyz"
Docs for reference: - Workers (Prefect 3.x) - Worker healthchecks If you share: - the pool name - how you started the temp worker (the exact command) - a short snippet of its logs …I can quickly confirm whether it’s picking up runs and suggest any fixes.