Skip to content

Latest commit

 

History

History

rightsize_compute_instances

README.md

Azure Rightsize Compute Instances

What It Does

This policy checks all the instances in Azure Subscriptions for the average or maximum CPU and/or memory usage over a user-specified number of days. If the usage is less than the user provided Idle Instance CPU and/or memory percentage threshold then the Virtual Machine is recommended for deletion. If the usage is less than the user provided Underutilized Instance CPU and/or Memory percentage threshold, and the total number of currently attached disks is supported with a smaller instance size, then the Virtual Machine is recommended for downsizing. Both sets of Virtual Machines returned from this policy are emailed to the user.

How It Works

  • The policy leverages the Azure API to check all instances and then checks the instance average or maximum CPU and memory utilization over a user-specified number of days.
  • The policy identifies all instances that have CPU and/or memory utilization below the user-specified idle thresholds and provides the relevant recommendation.
  • The recommendation provided for Idle Instances is a deletion action. These instances can be deleted or powered off in an automated manner or after approval.
  • The policy identifies all instances that have CPU and/or memory utilization below the user-specified underutilized thresholds and provides the relevant recommendation.
  • The recommendation provided for Underutilized Instances is a downsize action. These instances can be downsized in an automated manner or after approval.

Policy Savings Details

The policy includes the estimated monthly savings. The estimated monthly savings is recognized for idle resources if the resource is terminated, and for underutilized resources if the resource is downsized.

  • The Estimated Monthly Savings is calculated by multiplying the amortized cost of the resource for 1 day, as found within Flexera CCO, by 30.44, which is the average number of days in a month.
  • For idle resources, the savings is the full cost of the resource. For underutilized resources, the savings is the difference of the current cost of the resource and the estimated cost of the recommended resource type.
  • Since the costs of individual resources are obtained from Flexera CCO, they will take into account any Flexera adjustment rules or cloud provider discounts present in the Flexera platform.
  • If the resource cannot be found in Flexera CCO, the Estimated Monthly Savings is 0.
  • The incident message detail includes the sum of each resource Estimated Monthly Savings as Potential Monthly Savings.
  • Both Estimated Monthly Savings and Potential Monthly Savings will be reported in the currency of the Flexera organization the policy is applied in.

Input Parameters

  • Email Addresses - Email addresses of the recipients you wish to notify when new incidents are created.
  • Azure Endpoint - The endpoint to send Azure API requests to. Recommended to leave this at default unless using this policy with Azure China.
  • Minimum Savings Threshold - Minimum potential savings required to generate a recommendation.
  • Allow/Deny Subscriptions - Determines whether the Allow/Deny Subscriptions List parameter functions as an allow list (only providing results for the listed subscriptions) or a deny list (providing results for all subscriptions except for the listed subscriptions).
  • Allow/Deny Subscriptions List - A list of allowed or denied Subscription IDs/names. If empty, no filtering will occur and recommendations will be produced for all subscriptions.
  • Allow/Deny Regions - Whether to treat Allow/Deny Regions List parameter as allow or deny list. Has no effect if Allow/Deny Regions List is left empty.
  • Allow/Deny Regions List - Filter results by region, either only allowing this list or denying it depending on how the above parameter is set. Leave blank to consider all the regions.
  • Exclude Stopped Virtual Machines - Whether or not to filter stopped virtual machines from the results. If set to Yes, only running virtual machines will be included in the results.
  • Exclude Databricks - Whether or not to filter virtual machines used for Azure Databricks from the results. If set to Yes, virtual machines for Azure Databricks will not be included in the results.
  • Idle Instance CPU Threshold (%) - The CPU threshold at which to consider an instance to be 'idle' and therefore be flagged for deletion. Set to -1 to ignore CPU utilization for idle instance recommendations.
  • Idle Instance Memory Threshold (%) - The Memory threshold at which to consider an instance to be 'idle' and therefore be flagged for deletion. Set to -1 to ignore memory utilization for idle instance recommendations.
  • Underutilized Instance CPU Threshold (%) - The CPU threshold at which to consider an instance to be 'underutilized' and therefore be flagged for downsizing. Set to -1 to ignore CPU utilization for underutilized instance recommendations.
  • Underutilized Instance Memory Threshold (%) - The Memory threshold at which to consider an instance to be 'underutilized' and therefore be flagged for downsizing. Set to -1 to ignore memory utilization for underutilized instance recommendations.
  • Idle/Utilized for both CPU/Memory or either - Set whether an instance should be considered idle and/or underutilized only if both CPU and memory are under the thresholds or if either CPU or memory are under. Has no effect if other parameters are configured such that only CPU or memory is being considered.
  • Threshold Statistic - Statistic to use when determining if an instance is idle/underutilized.
  • Statistic Interval - The interval to use when gathering Azure metrics data. Smaller intervals produce more accurate results at the expense of policy memory usage and completion time due to larger data sets.
  • Statistic Lookback Period - How many days back to look at CPU and/or memory data for instances. This value cannot be set higher than 90 because Azure does not retain metrics for longer than 90 days.
  • Exclusion Tags - The policy will filter resources containing the specified tags from the results. The following formats are supported:
    • Key - Filter all resources with the specified tag key.
    • Key==Value - Filter all resources with the specified tag key:value pair.
    • Key!=Value - Filter all resources missing the specified tag key:value pair. This will also filter all resources missing the specified tag key.
    • Key=~/Regex/ - Filter all resources where the value for the specified key matches the specified regex string.
    • Key!~/Regex/ - Filter all resources where the value for the specified key does not match the specified regex string. This will also filter all resources missing the specified tag key.
  • Exclusion Tags: Any / All - Whether to filter instances containing any of the specified tags or only those that contain all of them. Only applicable if more than one value is entered in the Exclusion Tags field.
  • Skip Instance Sizes - Whether to recommend downsizing multiple sizes. When set to 'No', only the next smaller size will ever be recommended for downsizing. When set to 'Yes', more aggressive downsizing recommendations will be made when appropriate.
  • Automatic Actions - When this value is set, this policy will automatically take the selected action(s).
  • Power Off Type - Whether to perform a graceful shutdown or a forced shutdown when powering off idle instances. Only applicable when taking action against instances.

Please note that the "Automatic Actions" parameter contains a list of action(s) that can be performed on the resources. When it is selected, the policy will automatically execute the corresponding action on the data that failed the checks, post incident generation. Please leave it blank for manual action. For example if a user selects the "Delete Instances" action while applying the policy, all the resources that didn't satisfy the policy condition will be deleted.

Policy Actions

  • Sends an email notification
  • Power off virtual machines (if idle) after approval
  • Delete virtual machines (if idle) after approval
  • Downsize virtual machines (if underutilized) after approval

Prerequisites

This Policy Template uses Credentials for authenticating to datasources -- in order to apply this policy you must have a Credential registered in the system that is compatible with this policy. If there are no Credentials listed when you apply the policy, please contact your Flexera Org Admin and ask them to register a Credential that is compatible with this policy. The information below should be consulted when creating the credential(s).

Credential configuration

For administrators creating and managing credentials to use with this policy, the following information is needed:

  • Azure Resource Manager Credential (provider=azure_rm) which has the following permissions:

    • Microsoft.Insights/metrics/read
    • Microsoft.Compute/skus/read
    • Microsoft.Compute/locations/vmSizes/read
    • Microsoft.Compute/virtualMachines/read
    • Microsoft.Compute/virtualMachines/write*
    • Microsoft.Compute/virtualMachines/powerOff/action*
    • Microsoft.Compute/virtualMachines/start/action*
    • Microsoft.Compute/virtualMachines/delete*

    * Only required for taking action; the policy will still function in a read-only capacity without these permissions.

  • Flexera Credential (provider=flexera) which has the following roles:

    • billing_center_viewer

The Provider-Specific Credentials page in the docs has detailed instructions for setting up Credentials for the most common providers.

Supported Clouds

  • Azure

Cost

This Policy Template does not incur any cloud costs