Need some troubleshooting help with a discrepancy in our Architect flow performance metrics. A Data Action calling our on-premise CRM via a secure tunnel occasionally returns an HTTP 504 Gateway Timeout from the proxy layer. However, the Genesys Cloud Performance Dashboard logs these as ‘Flow Execution Timeout’ rather than ‘External Service Error’. This distinction is critical for our SLA reporting. Is there a configuration setting to ensure the root cause HTTP status code is preserved in the dashboard metrics instead of being categorized as a generic platform timeout?
The problem here is…
- The dashboard aggregates based on the Data Action’s internal timeout threshold, not the raw HTTP response. If the action exceeds its configured
timeout(default 10s), Genesys Cloud kills the request and logs a generic execution failure. The 504 might happen after the local timeout, or the platform simply categorizes any non-2xx response beyond a certain duration as a flow error. - Check the Data Action configuration. Ensure the timeout value is significantly higher than the expected latency plus proxy overhead.
- Enable verbose logging for the specific flow version. Look at the raw JSON payload in the interaction logs. The
errorobject often contains the underlying HTTP status code even if the dashboard summary is generic. - If using Terraform for the flow definition, inspect the
genesyscloud_flowresource. There isn’t a specific flag to force HTTP code preservation in the summary metrics, but you can export interaction data to S3/Azure Blob via the Analytics API. A custom script parsing the raw interaction JSON will give you the exacthttp_status_codefield, bypassing the dashboard’s aggregation logic entirely.