How do I correctly to configure BYOC edge WebSocket connections for high concurrent load testing? We are running a performance test using JMeter 5.6.2 to simulate 2,000 concurrent agents connecting to a Genesys Cloud BYOC edge instance in the US-East region. The test scenario involves each virtual agent establishing a WebSocket connection to the /v2/communications endpoint and maintaining the connection for 30 minutes to simulate active call sessions. After approximately 15 minutes, we start seeing a significant increase in connection failures. The JMeter logs show WebSocket connection failed: 1006 Abnormal Closure for roughly 15% of the virtual users. The BYOC edge logs indicate that the maximum number of concurrent WebSocket connections has been reached, but the specific limit value is not clearly documented in the current admin console. We have verified that the network firewall is not dropping packets and that the TLS 1.2 handshake is successful before the failure occurs. The error does not seem to be related to API rate limiting on the REST endpoints, as those continue to return 200 OK for configuration updates. The issue appears isolated to the persistent WebSocket layer required for real-time media signaling. We are using the default BYOC edge configuration provided by Genesys Cloud, with no custom modifications to the WebSocket backlog or thread pool sizes. The load test is executed from a dedicated EC2 instance in us-east-1 to minimize network latency. We need to understand if there is a hard limit on concurrent WebSocket connections per BYOC edge node and how to scale this limit to support our target volume of 5,000 concurrent agents. Is there a specific configuration parameter in the BYOC edge settings that controls the maximum number of simultaneous WebSocket connections? We have checked the max_connections setting in the edge configuration file, but it is not clear if this applies to WebSocket connections or just general TCP connections. The documentation mentions scaling the edge cluster, but we want to optimize the single node capacity first. Any guidance on the expected WebSocket connection limits and how to adjust them for high-density load testing scenarios would be appreciated. We are currently unable to proceed with the full-scale performance validation due to these connection drops.