Does anyone know if there is a hard limit on simultaneous websocket connections per tenant for the Platform API? Running a basic JMeter script hitting /api/v2/conversations to simulate 500 concurrent inbound calls. The test works fine up to 150 users. After that, the backend starts returning 1001 close codes randomly.
The environment is Genesys Cloud EU1. Using the standard REST API to create the conversations, then switching to websocket for real-time updates. The documentation here mentions rate limiting for REST, but nothing specific about websocket connection caps for load testing scenarios.
Is this a known bottleneck for the platform_api tier? Or is the issue likely on the client side with how the websocket heartbeats are handled under load? The server logs just show clean disconnects, no 5xx errors. Need to figure out if we are hitting a gateway limit or if the test setup is flawed. Any insights on scaling websocket connections for performance validation?