Hi all - having major difficulties working out how to ensure no cached image in LiveKit Cloud. Some mystery code that absolutely doesn't live in my codebase anymore is still executing. I even deleted & re-created the agents. Possibilities: 1. LiveKit Cloud has server-side caching that ignores new deployments (unlikely...) 2. Old agent instances from deleted agents are still being routed to somehow (same agent name... but different ID - don't think this could be it?) 3. Docker layer cache on LiveKit's build servers contains old Python bytecode 4. LiveKit Cloud warm instances not being replaced when new images are pushed Can't think of anything else... any ideas?

you can kill their API key?

<https://cloud.livekit.io/projects/p_/settings/keys>

Not sure if you were asking about the API key or how to stop the process on the machine. If the it is a question of how to stop the "rogue" agent on the machine you can kill the process there using

ps

to find it or task manager on windows.

I tried to kill the API key but their process was still running and due to the agent name being the same there was a race condition between the cloud agent & local agent. I couldn't stop it no matter what I tried. The problem was my colleague wasn't available so I lost a lot of time!