Percent of Devices with WinML Runtime Error
Description
This measure monitors the overall health & reliability of Windows Machine learning feature.
The main intent of this measure is to track the percentage of devices experiencing Winml Runtime Error.
The measure uses only WinML Runtime Error telemetry and WinML SessionCreation telemetry.
Measure attributes
Attribute | Value |
---|---|
Audience | Standard |
Time period | 30 days |
Measurement criteria | Machine |
Minimum population | 30 devices, using confidence intervals |
Passing criteria | Less than 1% |
Measure ID | 32436662 |
Calculation
Using daily distinct devices having winml runtime error divided by daily distinct device count that used winml
Measure value = Daily Distinct DeviceId count that sent Winml Runtime error telemetry / Daily Distinct DeviceId count that sent Winml ProcessInfo telemetry