Need some help troubleshooting SIP trunk registration failures during a load test.
We are running a JMeter 5.6 script to simulate 500 concurrent SIP REGISTER requests against a Genesys Cloud US1 instance. The goal is to validate capacity before a peak season rollout.
The test starts fine, but after 100 concurrent connections, we start seeing HTTP 408 Request Timeout errors on the SIP signaling path. The JMeter thread group is set to a 10-second ramp-up, which should be well within limits, but the Genesys Cloud SIP endpoints seem to drop the connection before the 200 OK is returned.
Environment details:
- Genesys Cloud US1
- SIP Trunk configured with TLS 1.2
- JMeter 5.6 with HTTP(S) Test Script Recorder
- No custom Architect flows involved, just raw SIP registration
Is there a specific rate limit for SIP REGISTER requests that is lower than the general API limits? Or is this a WebSocket connection issue masquerading as a SIP error?
Any insights on how to tune the JMeter script or Genesys Cloud settings to handle this concurrency would be appreciated.