Thanks a lot for looking into this. Unfortunately the HPO is still not working. Although the instance is starting now we are receiving the error below in the console of the application instance. When manually enqueueing the same experiment it does work, leading us to believe that this is not an environment issue. It also happens for cloned runs that worked in the past. Could this be related to the previous issue?

Traceback (most recent call last):
  File "/usr/local/lib/python3.11/site...

Also, it does not kill the instance, it continues to send progress reports every 2 minutes. I added the full log in the attachment.

Hey John, I'm Nathans colleague, thanks for looking into this. It seems the application is starting now. What was the issue in the end?