Azure Alerts with ML

Hello World, 

Say hello to the Public Preview of the new and advanced Dynamic alerts in Azure.

Yes, Dynamic!

For a monitoring guy, I’m more than happy to see this very long-awaited feature.

So what it is?

In the OLD world we made a very static decision about the threshold, but in the modern world, we can work with simple anomaly detection. 

This DOC is providing more information about the new Dynamic alert.

How does it work?

Metric data collected by Azure been calculated and by a simple algorithm provides anomaly detection by using machine learning methods.

How I change the sensitivity?

The new logic asks you for the Alert Sensitivity parameter that can be High, Medium or Low. Every level will tighten itself to the series pattern and will generate a more or less false positive.

More than or Less than?

The alert condition is met when the threshold is above or below the pattern.

Number violations to trigger the alert

 Allows you to configure the minimum number of deviations required within a certain time window for the system to raise an alert.

 

2 Comments

Add a Comment

Your email address will not be published. Required fields are marked *