Skip to content

Call Quality Metrics KQL¶

Analyze call quality and identify common media performance issues.

Query Description¶

This query retrieves recent call diagnostic events, filters for poor quality, and summarizes the average latency, jitter, and packet loss for each call.

KQL Query¶

ACSCallDiagnosticsEvents
| where TimeGenerated > ago(1h)
| where MediaPathQuality != "Good"
| extend Latency = PacketLatencyMs, Jitter = PacketJitterMs, Loss = PacketLossRate
| summarize 
    AverageLatency = avg(Latency), 
    AverageJitter = avg(Jitter), 
    AverageLoss = avg(Loss) 
    by CallId, MediaPathQuality
| order by AverageLoss desc

Explanation¶

Field	Description
`TimeGenerated > ago(1h)`	Filters results to the last hour to focus on current issues and improve performance.
`MediaPathQuality != "Good"`	Selects only calls with poor quality or media drops.
`extend Latency, Jitter, Loss`	Renames the diagnostic metrics to more readable names.
`summarize AverageLatency, AverageJitter, AverageLoss`	Calculates the average values for each call ID and quality level.
`by CallId, MediaPathQuality`	Groups the results by call ID and media path quality.

Insights¶

Observed Latency: Look for average latency greater than 200ms, which may be noticeable to users.
Network Performance: High jitter or packet loss for a specific call ID suggests a network issue.
Volume Analysis: A high count of calls with poor quality may suggest a service-level issue or heavy load.

See Also¶

Sources¶

Azure Call Diagnostic Log Reference