Is it possible to isolate the latency contribution of external Data Actions when analyzing queue performance discrepancies in the Performance Dashboard?
Our EU-West BYOC environment utilizes complex Architect flows that invoke external Data Actions for real-time customer data enrichment prior to agent handoff. Recently, we have observed a significant variance between the reported Average Handle Time (AHT) in the Performance Dashboard and the actual duration recorded in our internal logging systems. The discrepancy appears correlated with intermittent HTTP 504 Gateway Timeout errors returned by the external API endpoints during peak traffic hours. While the Architect flow is configured with a 30-second timeout for these Data Actions, the Performance Dashboard metrics seem to include the entire duration of the failed attempt in the queue wait time rather than attributing it to system error or external dependency failure. This misattribution skews agent productivity metrics and complicates root cause analysis for service level agreements. We require a method to distinguish between legitimate customer wait time and technical latency introduced by these integration points. Clarification on how the dashboard aggregates these specific timeout events would assist in aligning our reporting with actual operational performance.