Azure Classroom Series – 17/Sept/2020

Azure monitor metrics & Logs

• Application Monitoring Data
• Guest OS monitoring Data
• Resource Monitoring Data
• Subscription Monitoring Data
• Tenant Monitoring Data

Metrics

• Metrics are collected at regular intervals & are identified by

  • name
  • timestamp
  • value

• Scenarios

  • Analyzing
  • Visualizing
  • Alerting
  • Automating
  • Exporting
  • Retrieving
  • Archiving

• Four Sources of Metrics

  • Platform:
    • Which we get without any configuration
    • By default they are collected at one-minute frequency
  • Guest OS
    • To enable guest OS metrics for Windows Machine the Windows Diagnostic extension should be installed, For Linux machines the agents like Influx Data, Linux Diagnostics agent etc..
    • These metrics are collected from OS of the Azure VM.
  • Application:
    • These metrics are created by Application Insights
  • Custom Metrics:
    • These metrics are that we define manually use Custom Metrics API Refer Here and here for API

• Lets create a metric

  1. Navigate to Azure Monitor
  2. Lets create a metric around CPU Average (Select Scope at VM and select linux vm)
  3. Now login into linux vm and execute the following

  sudo apt-get update
  sudo apt-get install stress -y
  

  4. This stress tool can generate artificial load on our vm. Lets start stressing our system for 10 min using the following command

  stress --cpu 8 --io 4 --vm 2 --vm-bytes 128M --timeout 10m
  

  5. Create graphs for Percentage CPU and one more for Disk Read Operation/sec and write Operations/sec

• Percentage CPU is at 100%, Don’t we need to inform concerned team about it and take the necessary action

• How can i raise alerts and take necessary corrective actions (Automatically) to prevent application failures