Quick question, has anyone seen this weird error? with the performance dashboard metrics in our eu-west (paris) tenant. specifically, the ‘average wait time’ metric displayed in the queue performance view for our primary support queue is consistently showing values approximately 18% lower than the actual duration recorded in the architect flow execution logs.
environment details:
genesis cloud version: current (as of october 2023)
region: eu-west
flow version: v3.1 (architect)
the issue manifests during peak business hours (09:00-11:00 cet). when reviewing the conversation detail views, the timestamps indicate that agents are receiving the transfer after an average of 45 seconds of hold time. however, the performance dashboard aggregates these same interactions and reports an average wait time of roughly 37 seconds.
this discrepancy is problematic for our service level agreements, which are strictly bound by the dashboard metrics. we have verified that the clock settings on the agent desktops (chrome 118) are synchronized with the server time, and there is no significant latency in the network traces.
we attempted to isolate the variable by creating a test flow that bypasses the initial ivr menu and goes directly to the queue. the mismatch persists, though the percentage difference reduces to about 5%. this suggests the issue may be related to how the dashboard calculates ‘wait time’ when complex routing logic or data actions are involved prior to the queue entry.
has anyone observed similar deviations between the architect flow logs and the performance dashboard metrics? we are looking for a definitive explanation of how the dashboard computes these values versus the raw flow data, and whether there is a known configuration setting in the queue settings or the performance view that affects this calculation. any insights into the underlying logic would be appreciated, as we need to reconcile these figures for our monthly reporting.