Azure VM Metrics Collection

Products

• CloudHealth

Issue/Introduction

This article provides an overview of how Azure VM metrics are collected and how to enable guest OS metrics collection, and details about the collection process.

Resolution

What metrics does the platform collect?

Azure Monitor automatically collects metrics for your VM host when you create the VM. These metrics include:

• CPU utilization
• Network utilization
• Memory utilization (except for classic gov_cloud VMs)

To collect metrics from the guest operating system (OS) of the VM, you need to install an agent. Guest OS metrics can provide more detailed insights into your VM's performance.

Installing an Azure Monitor Agent for the Metrics Collection.

Refer to the following documentation for instructions on installing an agent to collect guest OS metrics:

https://docs.vmware.com/en/VMware-Tanzu-CloudHealth/SaaS/working-with-reports-recommendations-vmware-tanzu-cloudhealth/GUID-azure-rightsizing-new.html#enable-publishing-virtual-machine-metrics-to-azure-monitor-api-30

Metric storage tables

• CPU: azure_vm_cpu_utilization_*
• Network: azure_vm_network_utilization_*
• Memory: azure_vm_memory_utilization_* (except classic gov_cloud VMs)
• Disk (guest OS agent required): azure_vm_file_system_utilization_*

Metric collection process

• Metrics are collected for all VMs every 4 hours.
• Each collection job retrieves metrics for a specific time window (typically 4-8 hours prior to the job execution).
• During collection, the following VM information is retrieved:
  • Instance ID
  • Operating System (Windows or Linux)
  • VM type (ARM or Classic) - determines the appropriate metrics API call

Additional notes

• This document provides a general overview. Specific metrics available may vary depending on your VM configuration.
• For detailed information on available metrics and their definitions, refer to https://learn.microsoft.com/en-us/azure/azure-monitor/agents/agents-overview.