I deployed a Reasoning Engine to Agent Engine using the Vertex AI SDK. I deleted it with reasoning_engines.ReasoningEngine.delete() and confirmed via ReasoningEngine.list() that it’s gone.
But logs show it’s still running:
“Waiting for child process [xxxx]”
“Child process [xxxx] died”
The engine ID: xxxxxxxxxxx
Region: region
Project: project-id
It’s been looping for hours and I have no way to terminate it from the CLI, SDK, or console. I don’t have support plan access. Can a Googler escalate this to the backend team to shut it down?
I’ll upgrade support if necessary but I want to know if this is a bug or expected behavior.