Testing out, but getting no replies on 49100

Hello there,

I just started exploring the https://docs.nvidia.com/cloudxr-sdk/latest/samples/end_to_end_example.html, but for whatever reason I continue to get
? Stream stopped: Signaling connection failed (Error code: 0xC0F22202) when I try to connect to the LOVR plugin that was started with ./run.sh --webrtc from the simple webgl interface github.com/NVIDIA/cloudxr-js-samples/tree/main/simple.

As background, I'm running both the web client (cloudxr.js running via the docker method) and the lovr process on the same machine. It looks like something is going wrong when the websockets try to get started The connection to ws://localhost:49100/sign_in?peer_id=peer-3211025130&version=2&reconnect=1 was interrupted while the page was loading.

But on the LOVR end all I see is main: thread_init: already added for thread which indicates that some sort of connection got started briefly?

I tried using the other port for the web client with https://localhost:8443, but I think a bigger problem occurs when I try to communicate with the LOVR process on 49100 because the button Click https://localhost:49100/ to accept cert takes me to a not found page. I set up a caddy web proxy to make the cloud instance's domain name point back to the process running on port 49100, but it then gives me a page that says Error code: 501 Not Implemented.

Not sure what else to do for troubleshooting this. Is the core issue that I'm using a machine with an A100 on it instead of an rtx a6000 ada edition, or an l40/s?

Start with the CloudXR end-to-end example, the simple WebGL interface, and the LOVR process launched by ./run.sh --webrtc. Reproduce the websocket failure on port 49100 and compare the web client and LOVR output, including the signaling error and thread_init message. Done means the web client establishes a signaling connection to the LOVR process.