Anomalous Latency Alerts

Poor in-session responsiveness is the primary cause for poor session experience. The Anomalous Latency Alerts feature alerts administrators when there is a significant deviation in the session latency values. The proactive alerting helps administrators identify specific locations or Delivery Groups from which poor sessions might be originating.

A machine learning model is used to determine the baseline latency value for all Delivery Group-Location pairs for a specific customer. The baseline latency value is calibrated every day based on the ICARTT values from the last three days. Any outlier measurements of ICARTT are ignored. If the measured ICARTT has a deviation of 60% or more from the baseline latency value, an alert is generated.

For more information, see the Alerts article