Just noticed that our Premium App integration is hitting intermittent 504 Gateway Timeout errors when executing custom data actions via /api/v2/data/actions/{actionId} during high-concurrency Architect flows. The payload size remains under 10KB, and the underlying Lambda function completes in under 200ms, suggesting the timeout originates within the Genesys Cloud platform’s orchestration layer rather than our backend. This occurs specifically when more than 50 concurrent calls trigger the same data action within a 10-second window, despite our API rate limit headers indicating sufficient remaining quota. Is there a documented threshold for concurrent data action executions per organization that causes the platform to drop the request before it reaches the webhook?
Pretty sure the API throughput caps out around 50 concurrent requests per tenant for data actions. The platform enforces strict rate limits to prevent cascade failures during peak load.
Try adding a JMeter constant timer to space out requests. This prevents hitting the WebSocket connection limits and reduces 504 errors. Check the response headers for X-RateLimit-Remaining.