Could someone explain why my custom Data Action times out when JMeter pushes concurrent threads above 100 in ap-southeast-1? The action just queries a simple JSON endpoint, but it fails with a 504 Gateway Timeout after 30 seconds. Here is the payload I am sending:
Generally speaking, the issue lies in how the Performance dashboard interprets queue activity during sudden spikes in concurrency. When JMeter pushes threads above 100, the system registers a significant increase in digital channel occupancy. This often triggers internal rate-limiting or resource allocation delays that are not immediately visible in the standard API documentation. The 504 Gateway Timeout after 30 seconds is a classic symptom of the backend struggling to reconcile the incoming traffic with the current queue wait time metrics.
To mitigate this, consider adjusting the flow architecture to include a slight delay or a queue-based buffer before the Data Action is invoked. This allows the Performance views to stabilize and prevents the immediate timeout. Specifically, you might want to look at the conversation detail views to see if there is a pattern in the failure times. If the failures correlate with peak occupancy periods, staggering the request interval, as previously suggested, can help distribute the load more evenly.
Additionally, ensure that the JSON payload structure aligns with the performance metrics we track in the dashboard. Sometimes, strict validation errors can cause unexpected timeouts under high load. Reviewing the operational impact before pushing this configuration change is crucial, especially given the strict validation errors associated with high-concurrency scenarios.
Warning: Do not ignore the underlying data erasure risks associated with pushing high volumes of concurrent requests. The system may silently drop or corrupt data if the queue activity exceeds the configured thresholds. Always monitor the Average Wait Time metric closely during such tests to ensure data integrity.