Debug Self-Hosted LiveKit Audio

I have moved my audio-only conferencing app off LiveKit Cloud and onto my own VPS running the official LiveKit server build. Users can authenticate and show up in both the client UI and the LiveKit admin dashboard, yet no sound is exchanged once they toggle the microphone. All recommended ports are already open on the VPS, but I have not inspected the firewall layer in depth, so it is very likely a ufw/iptables or NAT traversal issue (STUN/TURN, ICE candidates, etc.).

What I need you to do
• SSH into the VPS, review LiveKit’s YAML config, and verify that UDP and TCP ranges match what the app expects.
• Audit and adjust any firewall rules that might still block media packets.
• Confirm STUN/TURN operation and, if required, install and wire up coturn with valid certs.
• Test end-to-end with two browser clients and prove that bidirectional audio flows reliably.
• Hand over a brief report of the exact changes so I can reproduce the setup on staging and future servers.

Acceptance criteria
1. Two or more users in the same room can hear each other without having to refresh or use workarounds.
2. No regressions in authentication or room listing.
3. Documented list of open ports, modified rules, and any TURN credentials added.

Stack highlights: LiveKit server (latest), WebRTC, coturn, ufw/iptables on an Ubuntu VPS.

If you’re comfortable tracing RTP/ICE traffic and tightening firewall rules without disrupting core services, your help will get the app back to full functionality.