Hi, i’m using helm to create self hosted prefect-worker for my prefect cloud, when i created the worker pool, it is working perfect.

helm install prefect-worker prefect/prefect-worker -f values.yaml --set-file worker.config.baseJobTemplate=base-job-template.json

once i have update my base-job-template.json with some new variables, the helm upgrade is completed successfully, and i can see my change on the configMap, but in the cloud UI, the worker pool is not updating at all. only after deleting the worker pool, the new one is being created with the relevant changes, but all my flows/deployments are not running, as they are not recognizing my worker pool (is the same name as the deleted one). @Marvin

From the information I've gathered, here's what I can suggest: 1. Updating Worker Pool Configuration: Prefect allows you to update worker pool configuration via the Prefect UI or the Prefect CLI commands. It does not explicitly state that using

helm upgrade

itself will do a live update of the worker pool within Prefect Cloud. 2. Helm Upgrade for Prefect Worker: It seems that although the

helm upgrade

command updates the ConfigMap, it might not trigger an update to the worker pool as the Prefect Cloud UI expects. To remedy this, I suggest checking the Kubernetes guide on how to properly use Helm to deploy and update a worker which is set up to run Prefect Flows in Kubernetes. 3. Worker Pool Not Recognized After Update: After deleting a worker pool, you may need to re-register or re-associate your flows/deployments with the new or updated worker pool, even if it has the same name. Make sure any environment variables (including API keys) are set correctly after an update. Check this GCP worker guide for setting the appropriate environment variables. 4. Missing Worker Pool Issues on GitHub: The GitHub issues search did not return relevant results, so it's possible this might be an uncommon issue or not widely reported. Please review the work pools configuration or the Kubernetes guide for more precise instructions. If problems persist, consider reaching out to Prefect Support or opening an issue in the GitHub repository for a more tailored solution. Remember, I would advise you with more enthusiasm if I thought it would make a difference.

We have followed the Kubernetes guide, so why the cloud UI is not updating automatically when running helm upgrade?