Edge BYOC Trunk Registration Timeout During High Concurrency Load Test

Could someone explain the registration behavior for BYOC trunks under stress?

  • Running a load test with JMeter to simulate 500 concurrent inbound calls via our BYOC trunk.
  • The edge gateway logs show successful SIP INVITEs initially, but after 200 calls, we see 503 Service Unavailable responses from the Genesys Cloud edge.
  • The error message specifically mentions Trunk registration limit exceeded or similar capacity constraints.
  • Our trunk is configured for 1000 concurrent sessions in the Edge configuration, so this should not be hitting a hard limit yet.
  • We are using the standard SIP trunk configuration with TLS encryption enabled.
  • The issue seems to correlate with the WebSocket connection pool on the edge server becoming saturated.
  • API calls to GET /api/v2/telephony/providers/edges show the edge status as online, but the call flow fails to route to the queue.
  • We suspect the edge might be dropping connections before the Genesys Cloud core processes them due to internal rate limiting on the edge side.
  • Is there a specific timeout or retry policy we need to adjust in the JMeter script or the Edge configuration to handle this burst?
  • Any insights on how the edge handles SIP re-registration during high load would be appreciated.