Does anyone know why a standard HTTP request node within a complex Architect flow is intermittently returning a 504 Gateway Timeout error? The flow in question handles high-volume outbound campaign routing and utilizes a sequence of decision nodes before hitting an external CRM update endpoint. The issue is not consistent; it occurs approximately 15% of the time during peak business hours in the Europe/Paris timezone. The timeout specifically happens at the HTTP request step, even though the external service provider confirms that their API response time is well under the configured 30-second timeout threshold on the node.
The environment is a standard Genesys Cloud Enterprise tenant with no recent changes to the network topology or firewall rules. The HTTP request node is configured to send a POST request with a JSON payload containing customer ID and interaction status. The response from the external system is expected to be a simple 200 OK with a small JSON body. We have verified that the payload size is minimal and does not exceed any known limits. The error log in the Conversation Detail view shows the failure at the specific node, but the Genesys Cloud platform logs do not provide additional context beyond the generic 504 error code. No other flows in the same tenant are experiencing similar issues, which suggests the problem is isolated to this specific integration or configuration.
We need to determine if this is a platform-level limitation regarding concurrent HTTP requests from Architect flows or if there is a specific configuration setting that needs adjustment. The business impact is significant as it results in failed CRM updates and requires manual reconciliation. We have attempted to increase the timeout duration on the node to 60 seconds, but the error persists at the same frequency. Any insights into potential causes or diagnostic steps to isolate the issue would be greatly appreciated. We are looking for a solution that does not involve rewriting the entire flow architecture.