Hey everyone, I’ve run into a really strange issue with the WebRTC softphone stability when simulating high concurrent call volumes.
The environment is Genesys Cloud US1. Using JMeter 5.6 to drive API calls for call initiation. The goal is to validate WebSocket connection limits and call capacity planning for a client scaling to 5,000 concurrent agents.
During the test, after 10 minutes of steady state with 500 concurrent threads, the softphone clients start dropping connections. The error appears in the browser console and the Genesys Cloud logs.
WebSocket connection failed: 1006 Abnormal Closure. Reconnection attempts failing with 408 Request Timeout.
The API rate limits seem fine, but the media servers are rejecting the WebRTC handshake. Is there a specific limit on concurrent WebRTC streams per tenant that is being hit? Or is this a timeout issue with the STUN/TURN servers under load?
Need to know if this is a configuration issue or a platform constraint. The JMeter config is standard for call initiation, but the media path seems to be the bottleneck. Looking for advice on how to tune the load or identify the exact failure point.