Hey there! That error

"failed to send job request {\"error\": \"no response from servers\"}"

you’re seeing in the LiveKit logs usually means the server is having trouble chatting with the agent workers. This can happen for a few reasons: • No Available Agent Workers: Double-check that your agent workers are up and running, and properly registered in your GKE cluster. • Network Connectivity Issues: Take a look at your firewall rules, Kubernetes network policies, and DNS settings to make sure the LiveKit server and agent pods can talk to each other smoothly. • Agent Registration Problems: Make sure the agents are successfully registering with the server and that their logs don’t show any errors. Also, the agent-side error

"unhandled websocket message Err(Io(Kind(UnexpectedEof))"

hints there might be a hiccup with WebSocket communication, which could also tie back to network or connectivity issues. Here’s what you can try to fix this: 1. Confirm that your agent workers are running and healthy. 2. Make sure all the necessary LiveKit ports are open and not being blocked by any firewall rules. 3. Check the agent logs for any registration or connectivity issues. 4. Review your Kubernetes network policies to ensure they allow communication between the different parts. Hopefully, these tips help you get things back on track! If you want, I’m here to help with anything else you need. Sources: pkg/agent/client.go | Deploy to Kubernetes | LiveKit Docs | Remote method calls (RPC) | LiveKit Docs