Context:
Running a load test for Genesys Cloud Outbound Campaign API using JMeter 5.6.2 from our Singapore staging environment. The goal is to validate throughput for a campaign expecting 200 concurrent outbound calls per minute. We are simulating the initial SIP INVITE requests via the POST /api/v2/outbound/campaigns/{campaignId}/predictive endpoint to trigger calls.
The test setup involves a thread group with 50 users, ramp-up time of 10 seconds, and a loop count of 10. We are using the HTTP Request sampler to send the JSON payload containing the contact list and dialer settings. The environment is Genesys Cloud EU-West-1. We have verified that the API credentials have the necessary outbound:campaign:view and outbound:campaign:edit permissions. The network latency from Singapore to EU-West-1 is around 180ms, which should be acceptable.
Error:
After approximately 30 seconds into the test, we start seeing a significant increase in 408 Request Timeout errors. The JMeter response message shows:
408 Request Timeout
The server timed out waiting for the request.
The error rate spikes to about 40% when the concurrent thread count exceeds 30. We have tried increasing the httpclient.connect.timeout and httpclient.request.timeout properties in JMeter to 30000ms, but the issue persists. The server seems to drop the connection before returning a proper response.
We have also checked the Genesys Cloud activity logs and see no corresponding error entries, which suggests the request might not be reaching the application layer.
Question:
What is the correct way to handle SIP 408 Request Timeout errors during high-concurrency outbound campaign simulations? Are there specific API rate limits or connection pool settings in JMeter that need to be adjusted for Genesys Cloud Outbound APIs? We are looking for best practices to optimize the JMeter script and avoid these timeouts. Any insights on the expected throughput limits for the predictive dialer API would be helpful.