Context:
Trying to understand why our hosted AppFoundry integration is experiencing intermittent WebRTC connection drops when agents switch between two distinct Genesys Cloud organizations using multi-org OAuth. The SDK version is 1.8.2. Error logs show a 403 Forbidden on the media endpoint after the initial SIP invite succeeds.
Question:
Is there a known limitation with WebSocket persistence across tenant contexts? We are seeing the signaling work, but the media stream fails to establish after the token refresh cycle completes.
You should probably look at at the iceTransportPolicy settings in the SDK config. Multi-tenant OAuth often triggers a re-auth handshake that resets the media socket. Setting forceRelay: true and ensuring iceCandidatePoolSize is at least 10 usually prevents those 403s during context switches.
You should probably look at at the risk of forcing relay. While forceRelay: true stabilizes the connection, it introduces significant latency and increases bandwidth costs for the enterprise.
Have you tried adding a constant timer of 1000ms in your JMeter script?
The 403s usually hit when the WebSocket reconnects faster than the platform can validate the new multi-org token. Slowing down the handshake rate in the load test often reveals if it’s a client-side race condition rather than a SIP issue.