AIOps with Azure Metrics Advisor

How to get started with Azure Metrics Advisor

What is Azure Metrics Advisor

Azure Metrics Advisor is an Applied AI Service designed to protect the performance of your organization’s growth engines, such as key IT services and operations. Its AI-powered monitoring features will not only allow you to stay one step ahead of incidents, but also assist you to quickly troubleshoot the probable cause of a problem.

Data preparation: map your AIOps problem to the dataset

The first key to successfully solving your AIOps problem is to format your dataset in a way that Azure Metrics Advisor (AMA) can digest. To help you get familiar with the data schema required for AMA to deliver multi-dimensional anomaly detection (AD) and root-cause analysis (RCA) for your business and service, let’s take an example.

  • Metric ‘DAU (Daily Active Users)’, which indicates customer engagement status
  • Metrics ‘PLT (Page load time)’ and ‘CHR (Cache Hit Rate)’, which track their website service running status.

Data onboarding

After all the metrics are ready in the above format, the next step is to onboard the formatted dataset to Azure Metrics Advisor. This is a simple step to register parameters such as what the data source is, how to connect, how to filter and aggregate, and what is the ingestion granularity (frequency). The entire onboarding process will only take a few minutes if you have the parameters handy. For more details, you can also refer to Azure Metrics Advisor documentation: Onboard your data feed to Metrics Advisor — Azure Cognitive Services | Microsoft Docs.

Tuning the detection

To optimize the anomaly detection results based on your business logic and contexts, there are two ways to customize your anomaly criteria:

  • Metric level — tuning at this level will change all time-series under that metrics
  • Group level — tuning at this level will change all time-series under that group, e.g., a group of specific dimension values (category or region in the above table)
  • Series level — tuning at this level will change only one time series
  • Adaptive learning: you can provide feedback on which data points should have been detected as anomalies and which ones should have been marked as normal. Azure Metrics Advisor will automatically learn from your feedback for future detections through underlying reinforcement learning algorithms.

Root cause analysis

In addition to the detected anomalies, Azure Metrics Advisor also offers insights into what might have been the cause of the issue to help you further troubleshoot. Stakeholders can get detection results via communication channels set up by the user, along with the anomaly detected and root cause analysis.

Reference architecture

Finally, let’s see how Contoso integrated AMA in their end-to-end AIOps workflow and architecture to monitor their business and IT service KPIs.

  • Configure and fine-tune the anomaly detection model
  • Identify and correlate anomalies
  • Diagnose anomalies and help with root cause
  • Configure alerts for each metric
  • Enable integration with the workflow engine via webhook

Get started today

Go to the Azure portal to create your new Metrics Advisor resource here. You can also read the Metrics Advisor document to learn more about the service capabilities. And you can find the sample dataset mentioned in the blog here.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store