Quick question about HTTP action timeouts when calling a Node.js backend hosted on AWS ECS from a Genesys Cloud Architect flow. The integration is deployed via AppFoundry as a Premium App with proper CORS headers and OAuth 2.0 client credentials flow configured for the /api/v2/integrations/webhooks endpoint. During peak load testing using JMeter with 50 concurrent sessions, approximately 15% of requests fail with a 502 Bad Gateway error originating from the Genesys Cloud edge. The backend logs show successful 200 OK responses within 800ms, well under the default 30-second timeout, suggesting the issue lies within the Genesys Cloud platform’s connection handling or proxy layer rather than the application logic. The error message returned in the Architect trace is ‘Connection reset by peer’ despite valid TLS 1.2 certificates. Has anyone encountered this specific behavior with high-concurrency webhook callbacks? Adjusting the retry policy in Architect does not seem to mitigate the initial failure rate, and increasing the timeout duration has no effect since the backend responds promptly. The environment is US-East-1 with standard AppFoundry networking configurations. Need to determine if this is a known platform limitation or a misconfiguration in the webhook registration.