Reclaiming the unused and overallocated Azure virtual machines

Use the out-of-the-box Azure views to identify the unused and overallocated virtual machines that are provisioned in the Azure infrastructure. The views also provide recommendations to reclaim these VMs for optimizing the capacity and improving the efficiency of your Azure infrastructure.

For more information, see the following sections:


Reference information

 About Azure views

Azure views enable you to view and manage the capacity of Azure virtual machines in your business service environment. They provide an overview of available capacity, utilization, trends, and residual capacity in terms of CPU, memory, and datastore utilization for each entity. 

Azure views also provide a set of recommended actions to help you resolve issues that are identified in your Azure cloud environment.

 Quick definitions (Idle VMs and Overallocated VMs)

TermDefinition
Unused or idle VMs
  • VMs that are currently powered off and were powered off for a significant period in the last 30 days
  • VMs with low CPU utilization in the last 30 days
Overallocated VMs

VMs with low utilization of the resources (CPU, Memory, or Storage) that were allocated to them

 About recommendations

A recommendation is a suggestion based on best practices about how to avoid potential capacity risks or improve efficiency. Recommendations can be of two types: Risk and Efficiency.

  • Risk recommendations are provided to prevent a risk. For example, in case a disk is nearing saturation; the recommendation provides information about how much space must be added to prevent saturation for the next three months. 
  • Efficiency recommendations are not to prevent a problem, but to improve the overall efficiency of the data center and to avoid wastage of resources and money. For example, most of the old snapshots consume resources to save images that are not used for a long time. The recommendation provides information about the saving that you can achieve by deleting these snapshots to save resources and improve the data center efficiency.

For more information about the conditions and associated recommendations, see Using recommendations to resolve capacity risk and efficiency issues.

Scenario

Nathan is an Azure Technology Specialist in an IT organization. He manages and monitors the Azure infrastructure to ensure that the existing infrastructure capacity has sufficient resources (CPU, memory, and storage) to meet the current and future demands.

Current problem: Nathan wants to know whether the current Azure infrastructure elements are optimally and efficiently utilized. He wants to identify whether there are any overprovisioned or unused Azure VMs and take necessary actions on such infrastructure elements.

Solution: Nathan can use the Recommendations page in the Azure views to quickly and easily get the information that he wants. The Recommendations page in the Azure Virtual Machines view provides a list of elements that have overallocated and idle VMs and recommended actions to manage them.

Prerequisites: Let us assume that Alan, the TrueSight Capacity Optimization administrator, has already completed the necessary prerequisite tasks to set up the Azure views so that Nathan can view and use the Azure views.

Workflow 

The following illustration provides a high-level view of the process.

AWS

Steps

Nathan performs the following steps:

  1. Open the Azure Virtual Machines view

    1. Log on to the TrueSight console.
    2. In the left navigation pane, click Capacity > Views > Cloud > Azure > Virtual Machines.
       

  1. View the VM-specific core metric details

    On the Virtual Machines page, review the following column values (metrics) to understand the utilization of the key resources:

    ResourceCPUMemoryDescription
    MetricsCPUMemoryShows the total resource value for the Azure virtual machine.
    CPU UtilMemory UtilShows the current utilized resource value.
    CPU TrendMemory TrendShows the rate at which the resource utilization is increasing or decreasing or is staying stable.

    By default, some columns are hidden. Use the action menu that is located next to the table title to display the hidden columns.

  2. Configure the optimization behavior settings

    1. On the Virtual Machines page, select   > Edit Settings.
    2. In the Other Settings, configure the optimization behavior settings.
       Click here to view the details of the optimization behavior settings

      Optimization behaviorDescription
      Aggressive

      Resource utilization in the server is computed by considering the average value of hourly samples. Spikes in the resource utilization within the hour are not considered. Then, 95th percentile of the hourly value over the last 30 days is computed for each resource to generate the configuration of the target virtual machine or instance type.

      Balanced (Default)

      Resource utilization in the server is computed by considering the 90th percentile value of hourly samples. Only 10% of spikes in the utilization within the hour are not considered. Then, 95th percentile of the hourly value over the last 30 days is computed for each resource to generate the configuration of the target virtual machine or instance type.

      Conservative

      Resource utilization in the server is computed by considering the 99th percentile value of hourly samples. Only 1% of spikes in the utilization within the hour are not considered. Then, 95th percentile of the hourly value over the last 30 days is computed for each resource to generate the configuration of the target virtual machine or instance type.

      Information

      If you select the Conservative or Balanced optimization behavior for a server that is not instrumented, the results are based on the Aggressive behavior.

      BMC recommends the Conservative behavior for servers that are running business-critical applications.

  3. View the recommendations for the overallocated and idle VMs

    1. Click the Recommendations tab. 
      The Recommendations page displays a table that lists overallocated and idle VMs with efficiency and risk related issues, and recommendations to fix them.

      Information

      If TrueSight Cloud Cost Control is installed, the Efficiency recommendations also provide the estimated cost savings (in $) that you can achieve by performing the recommended actions. The cost savings are computed based on the bill amount of your last or current month.

      Before you install TrueSight Cloud Cost Control, see License entitlements for TrueSight Cloud Cost Control .

    2. To view a list of overallocated VMs and idle VMs with either high or low efficiency only, click the High Efficiency or Low Efficiency quick filter buttons.

    3. To view more details of a VM, click the Severity icon corresponding to the VM.

      The details are displayed in a new window. Observe the information such as severity or efficiency (low or high), optimizer findings, and recommended actions.
      Consider the following examples with sample data:

      Idle VM with high efficiency

      Idle VM with low efficiency

      Overallocated VM with high efficiency

       Click here to review the recommendations details

      The following table explains the different fields that you see in the Recommendations window.

      FieldDescription
      SystemName of the virtual machine
      Optimizer findingsBrief description of the issue
      TypeRecommendation type. It is one of the configuration parameters in the optimizer rule that generates this recommendation
      Severity/EfficiencyEfficiency level of the recommendation that is based on the estimated savings per month that you can achieve by implementing the recommended action
      Generated onDate of the last run of the associated optimizer rule
      Optimizer ruleName of the associated optimizer rule
      Recommended actionsSuggested actions that you can implement to optimize your costs
      Accuracy indicator

      Displayed only for overallocated VM. Indicates the level of accuracy or reliability of the resizing recommendation. For more information about how the recommendation is generated, see Overallocated VM recommendation.

      You can click the icon to view its description.

      The icon can have one of the following values:

      Value - Accuracy level

      - Medium

      - High

      - Very High

      For more information about each accuracy level, see Accuracy levels of resizing recommendations for overallocated VMs.

      For more information about recommendations, see Using recommendations to resolve capacity risk and efficiency issues.

    4. Click x to close the window.

  4. Save an offline copy of the recommendations

    Click Settings > Export to Excel to export the recommendations to an Excel file for offline reference or to share it with other stakeholders. 

Was this page helpful? Yes No Submitting... Thank you

Comments