Running load tests for inbound SIP traffic using JMeter. At 200 concurrent sessions, getting 503 Service Unavailable responses from the Genesys Cloud SIP trunk. Lower volumes (50 concurrent) work fine. Is there a default concurrency cap for SIP trunks in the US1 region? Need to validate capacity before our Q4 peak. Any config tweaks or rate limit docs for SIP endpoints?
This 503 behavior typically stems from resource exhaustion in the SIP edge nodes rather than a hard-coded trunk limit. While Genesys Cloud scales dynamically, high-concurrency bursts like JMeter tests often exceed the default session allocation rate for a single trunk configuration.
Check the Trunk Configuration settings in Admin > Phone System. Specifically, verify the Maximum Concurrent Sessions value. If this is set to a static number (e.g., 100 or 200), it will reject new registrations once that threshold is hit, returning 503. Ensure this is set to “Unlimited” or a value significantly higher than your peak test load.
Additionally, review the Call Control settings for the associated DID ranges. If you are using number pools, ensure the Maximum Concurrent Calls setting on the Number Pool itself isn’t capping the flow.
For accurate capacity planning, I recommend submitting a support ticket referencing your specific Org ID. They can provide the current edge cluster capacity for your region (US1) and confirm if a capacity increase is needed for Q4. Relying on JMeter for definitive limits is risky due to how the load balancer distributes traffic across available SIP endpoints.