- JMeter 5.6.2
- Singapore datacenter
- Genesys Cloud Platform API
- Target: 500 concurrent WebSocket connections
What is the correct way to handle WebSocket connection limits when load testing an Architect IVR flow?
Trying to simulate high-volume inbound traffic using JMeter 5.6.2 from Singapore. The test plan initializes WebSocket connections to the platform API for call control. Single-thread tests work fine. Ramping up to 500 concurrent threads causes immediate connection failures. The logs show 502 Bad Gateway errors on the WebSocket upgrade request.
The error response is pretty generic. No specific rate limit header in the failure response. Just a blank body with status 502. The connections are dropping before any actual IVR logic executes. This happens even with a slow ramp-up time of 60 seconds.
Is there a documented limit on concurrent WebSocket connections per tenant or per IP? The platform docs mention API rate limits for REST endpoints but nothing specific for WebSocket throughput. Need to know if this is a hard limit or a configuration issue. Any suggestions for structuring the JMeter test plan to avoid hitting these limits? Trying to validate call capacity planning for a new deployment.