Why does this config cause our predictive outbound campaign to fail when we increase the concurrent agent count? We are running a load test to validate capacity planning for a new client onboarding. The environment is Genesys Cloud EU-West-1. We are using the latest Architect flow version with a simple predictive script.
At 10 concurrent agents, the system works perfectly. The dialer connects, the WebSocket streams data correctly, and the API calls to fetch customer records return 200 OK. However, when we scale up to 50 concurrent agents using JMeter to simulate the load, we start seeing a high volume of errors.
HTTP 429 Too Many Requests
The error occurs specifically on the /api/v2/predictivedialing/campaigns/{id}/sessions endpoint. The response body indicates that the rate limit has been exceeded. We have checked the API documentation and confirmed that our organization tier should support higher throughput. The JMeter configuration is set to a linear ramp-up over 60 seconds to avoid sudden spikes. Despite this, the Genesys Cloud platform seems to throttle the requests aggressively once the concurrent user count crosses the threshold of 30.
We have also verified that the WebSocket connection remains stable and does not drop. The issue is strictly related to the REST API calls made by the predictive dialer engine. The error rate increases proportionally with the number of concurrent agents. We are unable to determine if this is a bug in the predictive routing algorithm or a configuration issue on our side. The load pattern is consistent, with each agent handling an average of 20 calls per hour.
Has anyone else faced similar issues with predictive outbound campaigns at scale? We need to understand if there is a hidden rate limit per campaign or per organization that is not documented. We are also seeing increased latency in the response times, which affects the overall efficiency of the dialer. The JMeter logs show that the requests are being sent at a steady rate, but the responses are delayed or rejected. We are looking for guidance on how to optimize the configuration to handle higher concurrent loads without triggering these rate limits. Any insights would be appreciated.