Just noticed that the WebSocket connection for the Data Action API drops with a 1006 error code when the JMeter thread group hits 50 concurrent users. The environment is Genesys Cloud v2.98.0 in the US-East region. The test plan uses a simple JSON payload to trigger a custom data action named ‘LookupCustomerInfo’ via the /api/v2/analytics/events/realtime endpoint. The initial ramp-up phase with 10 threads works perfectly, returning 200 OK responses and valid JSON data. However, as soon as the concurrency increases to 50 threads per minute, the client-side logs show immediate connection closures without any retry logic from the SDK. The JMeter configuration sets the Content-Type to application/json and includes the necessary Bearer token for authentication. The timeout settings are set to 30000 milliseconds, which should be sufficient for a simple database lookup. The issue persists even when the data action logic is simplified to just a static JSON return, suggesting the problem lies in the connection handling rather than the backend processing time.
The goal is to validate the maximum throughput of the Data Action service before the peak season. The current setup uses the Genesys Cloud Java SDK version 14.5.0 for the WebSocket implementation. The error logs do not provide a specific reason for the 1006 close code, only that the server closed the connection unexpectedly. The network latency between the JMeter server and the Genesys Cloud endpoint is consistently under 50 milliseconds. The question is whether there is a specific concurrent connection limit for Data Actions that is lower than the standard API rate limits. The documentation mentions general WebSocket limits, but it does not specify if Data Actions share this pool with other real-time services like Agent Desktop or IVR. Any guidance on configuring the JMeter HTTP Request sampler to handle these disconnects gracefully or on identifying the exact bottleneck would be appreciated. The test environment is isolated, so no other traffic is interfering with the connections.