Skip to content

Azure Virtual Machine

Monitor your Azure Virtual Machines with Azure Monitor

Section titled “Monitor your Azure Virtual Machines with Azure Monitor”

Source: Monitor your Azure virtual machines with Azure Monitor

  • Virtual machines (VM) are hosting business services and we need to monitor to ensure they can host effectively without causing unnecessary costs.
  • To ensure availability of business services, respond to access, security, and performance issues.

Monitoring features:

  • Azure Monitor Metrics - collected to describe things like performance, utilization, errors, users
  • Azure Monitor Logs - system events with timestamp and data, available at Azure resource level. Stored in log analytics workspace

VM Monitoring Layers

  • Host - compute, storage, network, state, disk
    • Can trigger alerts, use for pattern analysis, cost control
  • OS
  • Client workloads
  • Applications in VM

OS and below monitoring requires Azure Monitor Agent.

  • Create a VM in portal and check it’s built in monitoring features
    • Create a Linux VM with alert and boot diagnostics
    • Start VM and see collection of basic metrics and activity logs
    • View built in metrics graphs, activity logs, boot diagnostics
  1. Use Metrics Explorer to view detailed host metrics

    • Metrics explorer has charts to display VM CPU, data and other metrics
    • Add metrics in the explorer and change types of charts and aggregation type like Avg, Min, Max
  2. Collect client performance counters by using VM insights

    • To monitor operating system workloads and applications, the Azure Monitor Agent needs to be installed inside the VM
    • VM insights:
      • Installs Azure Monitor Agent on your VM.
      • Creates a data collection rule (DCR) that collects and sends a predefined set of client performance data to a Log Analytics workspace
      • Presents the data in workbooks
  3. Collect VM client event logs

    • Set up a data collection rule (DCR) to collect Linux VM syslog data and view log data in log analytics using Kusto Query Language (KQL)
    • Created DCRs can select from different performance counters and sampling rates or custom counters