Azure Monitor AIOps alerts with Dynamic Thresholds is now generally available

Posted on Thursday, May 30, 2019

Metric Alerts with Dynamic Thresholds is now generally available. Since the public preview many customers have created alert rules that can detect and consider metric patterns, apply at scale to multiple dimensions or resources and are using high-level concepts. Get more in-depth information in the previous blog post or in the documentation.

In this release, we have incorporated a lot of customer feedback to further enhance the feature experience:

  • Enhanced experience for authoring alert rules
    • We simplified the metric alert rule authoring experience and now have a seamless experience, which will be easier for users who authored static threshold alert rules in the past.
    • We added recommendation info balloons that inform you about patterns we detect and scenarios in which Dynamic Thresholds could reduce noise and management overhead at scale.
  • Enhanced experience for fired alerts
    • The alerts view was revamped and now includes a chart showing the metric behavior at the alert time. This chart displays either the static threshold, or the dynamic thresholds used when the alert was fired.
    • Users can now provide feedback on the accuracy and experience of dynamic thresholds alerts. This feedback will impact the machine learning algorithmic research and improve future detections. 
  • Support for metric alerts rules for logs that use Dynamic Thresholds

For pricing information, please visit our pricing page. To get started, please refer to the documentation.

We would love to hear your feedback – If you have any questions or suggestions, please reach out to us at azurealertsfeedback@microsoft.com