I am currently performing a stress test on our primary Architect inbound flow. We are simulating a ‘Burst’ of one thousand concurrent calls arriving at the IVR within a five-second window. We are seeing a significant increase in ‘Flow Errors’ and some calls are being dropped before they even reach the first decision block. Does Genesys Cloud have a hard limit on the number of concurrent flow executions per organization, or is this a result of our custom Data Actions hitting their own rate limits?
Hello Sat60. I am a Senior Architect with twelve years of experience and I have seen many large-scale rollouts. There is no hard limit on concurrent flow executions for a single org, as the platform is designed to scale horizontally. However, every flow execution consumes resources. If your flow starts with a complex ‘Data Action’ or a ‘Bot Flow’ lookup, that is where the bottleneck will be. You should check the ‘Rate Limit’ metrics in the Developer Center to see if your Data Actions are returning 429 errors during the burst.
we have strict GDPR rules for our data handling. To follow up on Cam45, you should also look at your ‘Edge’ capacity if you are using BYOC Cloud or on-premise Edges. Even though the flow runs in the cloud, the signaling still has to be processed by the Edges. If your Edges are hitting their concurrent call limit, they will start rejecting calls before the Architect flow even has a chance to run. Make sure your Edge Group is properly sized for your peak burst traffic!