Looking for advice on SIP trunk registration failures observed during concurrent load testing.
- Running JMeter 5.6.2 with custom SIP sampler plugin to simulate inbound call volume.
- Target environment is Genesys Cloud pure cloud, US East region.
- Test scenario involves ramping up 500 concurrent SIP registrations within 60 seconds.
- Initial 100 registrations succeed with 200 OK response.
- Subsequent registrations fail with 408 Request Timeout after 10 seconds.
- Some requests return 403 Forbidden with message ‘Rate limit exceeded’.
- Checked API rate limits documentation, but SIP signaling seems separate from REST API limits.
- WebSocket connection pool in JMeter set to max 1000 connections.
- Network latency between test server and GC edge is under 5ms.
- No firewall blocks detected on port 5060 UDP/TCP.
- SIP trace shows REGISTER requests being sent correctly.
- Genesys Cloud admin portal shows no active incidents.
- Tried reducing ramp-up time, but issue persists at lower concurrency levels.
- Increased JMeter heap size to 4GB to avoid memory issues.
- Verified SIP trunk credentials are correct and not expired.
- Question is whether there is a hidden limit on SIP registration attempts per minute.
- Or if the load test configuration is triggering a security mechanism.
- Need to understand the exact threshold for SIP signaling rate limiting.
- Any guidance on how to properly scale SIP load tests without hitting these errors?
- Also interested in best practices for simulating high concurrent SIP traffic.
- Current setup uses single thread group for all registrations.
- Considering splitting load across multiple virtual users or thread groups.
- But first need to confirm if 403 is due to rate limiting or config error.
- Appreciate any insights from community members with similar experience.