Setting up a load test to validate our IVR capacity using JMeter. The goal is to simulate 200 concurrent callers hitting a single Architect flow that performs a database lookup via a Data Action (REST API) before routing.
When the concurrency hits ~50 threads, the Architect flow starts returning 504 Gateway Timeout errors on the Data Action node. Interestingly, the downstream API (hosted on AWS) shows no latency spikes and handles the requests fine in isolation. The Genesys Cloud platform seems to be dropping the connection or timing out the Data Action execution before the response returns.
Checked the platform logs, but nothing specific stands out other than the flow execution terminating. Is there a known limit on concurrent Data Action executions per flow or per organization? The standard documentation mentions general API rate limits, but this looks like a flow execution bottleneck rather than a simple HTTP 429.
Environment:
- Genesys Cloud US1
- Architect Flow: Simple Lookup → Route
- JMeter: 200 threads, 5s ramp-up
- Data Action: POST to internal microservice (avg latency 200ms)
Has anyone seen this behavior during scale testing? Adjusting the timeout on the Data Action node didn’t help. Need to understand if this is a hard limit on flow processing capacity.