Customizing Cloud Advisor

You can customize Cloud Advisor by postponing or dismissing recommendations and by customizing the logic that Cloud Advisor uses to make recommendations. Modify the recommendations based on your preferences and what works best for your workloads.

Customizing the Recommendation List

Cloud Advisor provides recommendations to help you maximize cost savings and improve the performance, security, and availability of your tenancy. However, you might want to customize recommendations for certain resources, postpone specific recommendations to review later, or dismiss recommendations entirely.

Recommendation overrides and changes to recommendation overrides will take effect the next time Cloud Advisor runs for your tenancy.

Customizing Overrides

To customize or override a recommendation for select resources

You can customize recommendations for resources in specific compartments and for resources with certain tags by creating recommendation overrides. You can create up to 50 recommendation customizations, including overrides. To create an override, you must specify at least one compartment or tag.

In addition to the standard Cloud Advisor permissions, to use overrides on compartments, you must belong to a group that has the COMPARTMENT_INSPECT permission. To use overrides with tags, you must belong to a group that has the TAG_NAMESPACE_INSPECT permission. For more information, see Details for IAM without Identity Domains.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Settings.

    The Cloud Advisor Settings page opens.

  2. In the Customizations and overrides section, under Overrides, click Create override.

  3. In the Create new override panel, enter the following information.
    1. For Name, enter a name for the override. The name must be unique.
    2. For Recommendation, in the drop-down menu, select the recommendation that the override applies to. The menu only lists recommendations that are customizable.
    3. In the Recommendation profile section, specify the recommendation profile to use for the resources impacted by this override.
      Important

      The Evaluation period and Methodology fields are only available for some profile types and might not apply.
      • For Evaluation period, select time span that you want Cloud Advisor to use to collect data for recommendations.
      • For Methodology, select either P95 or Average.
      • For Profile, select a profile. See descriptions of the profile options.
    4. In the Compartment overrides section, select up to 10 compartments for this override. The override applies to the resources in the selected compartments.
      1. For Specify compartments, select the method to use to identify the compartment: By name, By OCID, or By path.
      2. In the drop-down menu, select the compartments that the override applies to.
    5. In the Tag overrides section, click Add a tag override. The Add a tag override panel opens. Only defined tags are supported. The tag must be on the compartment, inherited from one compartment to another, or inherited from a compartment to the resource.
    6. In the Add a tag override panel, make the following selections, and then click Add. The override applies to the resources that contain the specified tags.
      1. For Tag namespace, in the drop-down menu, select the tag namespace that contains your tags. Cloud Advisor supports the following characters for the tag namespace field: 0-9, a-z, A-Z, _, @, -, :.
      2. For Tag key, in the drop-down menu, select the name used to refer to the tag. Cloud Advisor supports the following characters for the tag key field: 0-9, a-z, A-Z, _, @, -, :.
      3. For Value, select either Match any value or Match any of the following.
        • If you choose Match any value, the override applies to any resource with the tag selected in the Tag key field.
        • If you choose Match any of the following, in the drop-down menu, select or enter tag values. The override applies only to resources with the tag key and tag values specified. The tag values can contain any UTF-8 characters except single quotation marks.
    7. To add another tag override, click + another tag override and repeat the previous step. You can create up to 10 tag overrides.
  4. Click Create.
To edit a recommendation override

You can edit existing recommendation overrides.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Settings.

    The Cloud Advisor Settings page opens.

  2. In the Customizations and overrides section, in the Overrides table, find the recommendation that you want to edit, click the Actions icon (Actions icon), and then click Edit.
  3. In the Edit override panel, change the override, and then click Save changes.
To delete a recommendation override

You can delete recommendation overrides when they are no longer needed.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Settings.

    The Cloud Advisor Settings page opens.

  2. In the Customizations and overrides section, in the Overrides table, find the recommendation that you want to delete, click the Actions icon (Actions icon), and then click Delete.

  3. In the confirmation dialog, click Delete.

Customizing Recommendations for Specific Resources

To dismiss a recommendation for specific resources

You can dismiss recommendations so that the resources no longer appear in your recommendation list.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. Click the recommendation in the list to open the recommendation details panel. A list of individual resources with pending recommendations displays.
  3. Select the resources that you want to dismiss, and then click Dismiss selected.
  4. In the Dismiss recommendation panel, click Dismiss.
To postpone a recommendation for specific resources

You can postpone recommendations and set dates for them to reactivate.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. Click the recommendation in the list to open the recommendation details panel. A list of individual resources with pending recommendations displays.
  3. Select the resources for which you want to postpone recommendations, and then click Postpone selected.
  4. In the Postpone recommendations panel, for Select date, use the calendar picker to select a date when the recommendation reactivates, and then click Postpone.
To edit recommendation postponements for resources

You can change the date that a postponed recommendation reactivates.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. Click the recommendation in the list to open the recommendation details panel. A list of individual resources with pending recommendations displays.
  3. In the Cloud Advisor status menu, click Postponed. The list of resources whose recommendations have been postponed opens.
  4. Find the resources whose postponement you want to edit, click the Actions icon (Actions icon), and then click Edit postponement.
  5. In the Edit postponement panel, for Select date, use the calendar picker to select a date when the recommendation reactivates, and then click Save.
To reactivate a recommendation for specific resources

You can reactivate resources whose recommendations have been dismissed or postponed.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. Click the recommendation that has postponed or dismissed resources to open the recommendation details page.
  3. In the Cloud Advisor status menu, click Dismissed or Postponed, depending on whether the recommendation that you want to reactivate was dismissed or postponed.
  4. In the list, select the resources that you want to reactivate, and then click Reactivate selected.
  5. In the Reactivate selected resources panel, click Reactivate.

Customizing Recommendations for All Resources

To dismiss a recommendation for all resources

You can dismiss recommendations that you don't want to see in your dashboard.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. In the list, for the recommendation that you want to dismiss, click the Actions icon (Actions icon), and then click Dismiss.
  3. In the Dismiss recommendation panel, click Dismiss. The recommendation is dismissed for all associated resources.
To postpone a recommendation for all resources

You can postpone recommendations and set dates for them to reactivate.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. In the list, for the recommendation that you want to postpone, click the Actions icon (Actions icon), and then click Postpone.
  3. In the Postpone recommendation dialog, for Select date, use the calendar picker to select a date when the recommendation reactivates, and then click Postpone.
To reactivate a recommendation

You can reactivate recommendations that you have dismissed or postponed.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. In the Status menu, click Dismissed or Postponed, depending on whether the recommendation that you want to reactivate was dismissed or postponed.
  3. In the list, for the recommendation that you want to reactivate, click the Actions icon (Actions icon), and then click Reactivate.
  4. In the Reactivate recommendation panel, click Reactivate.

Filtering Recommendations

To filter the list of recommendations

You can filter the recommendation list based on category, service, and status.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. For Category, Service, and Status, select the desired filters.
To filter the resources in a recommendation list

For each recommendation, you can filter the resources that appear in the list by compartment, tag, and region. You can also save and share your filters.

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Recommendations.

    The Cloud Advisor Recommendations dashboard opens.

  2. Click the recommendation that you're interested in.
  3. In the Filtering menu, select Compartment, Tag, or Region.
  4. If you select Compartment, in the Filter compartments dialog, make the following selections, and then click Apply filter.
    1. Select the method to use to identify the compartment: By name, By OCID, or By path.
    2. In the drop-down menu, select the compartments that the override applies to.
    3. If you want to include subcompartments, select Include subcompartments.
  5. If you select Region, in the Filter region dialog, add regions, and then click Apply filter.
  6. If you select Tag, in the Filter tag dialog, make the following selections, and then click Apply filter.
    1. For Tag namespace, in the drop-down menu, select the tag namespace that contains your tags. Cloud Advisor supports the following characters for the tag namespace field: 0-9, a-z, A-Z, _, @, -, :.
    2. For Tag key, in the drop-down menu, select the name used to refer to the tag. Cloud Advisor supports the following characters for the tag key field: 0-9, a-z, A-Z, _, @, -, :.
    3. Select either Match any value or Match any of the following.
      • If you choose Match any value, the filter applies to any resource with the tag selected in the Tag key field.
      • If you choose Match any of the following, in the drop-down menu, select or enter tag values. The filter applies only to resources with the tag key and tag values specified. The tag values can contain any UTF-8 characters except single quotation marks.
  7. (Optional) To save your filter, in the Filtering menu, click Save filter settings. In the Save current filter set dialog, enter a name, and then click Save.

Customizing the Recommendation Profile

For some recommendation types, you can customize the logic that Cloud Advisor uses. The default configuration uses the standard profile and the average methodology.

Note

Changes made to your recommendation settings will take effect the next time Cloud Advisor runs for your tenancy.

Changing Your Profile Using the Console

  1. Open the navigation menu and click Governance & Administration. Under Cloud Advisor, click Settings.

    The Cloud Advisor Settings page opens.

  2. In the Customizations and overrides section, in the Global recommendations table, find the row for the recommendation that you want to customize, click the Actions icon (Actions icon), and then click Customize.

    A Customize recommendation logic panel opens.

  3. In the Customize recommendation logic panel, make the following selections (if applicable), and then click Save changes.
    Important

    The Evaluation period and Methodology fields are only available for some profile types and might not apply.
    • For Evaluation period, choose the time span that Cloud Advisor should use to collect data for recommendations.
    • For Methodology, select either P95 or Average.
    • For Profile, select a profile. See descriptions of the profile options.

    Changes made to your recommendation settings take effect the next time Cloud Advisor runs for your tenancy.

Changing Your Profile Using the CLI

To manage your recommendation profiles, use these API operations. For information about using the CLI, see Command Line Interface (CLI). For the full list of Cloud Advisor CLI commands, see Cloud Advisor CLI.

To get profiles, use the profile get command:

oci optimizer profile get --profile-id <PROFILE_OCID>

To create a profile, use the profile create command:

oci optimizer profile create --compartment-id, -c <COMPARTMENT_OCID> --description <TEXT> --levels-configuration <file://path/to/file>  --name <text>

To update profiles, use the profile update command:

oci optimizer profile update --profile-id <PROFILE_OCID>

To delete profiles, use the profile delete command:

oci optimizer profile delete --profile-id <PROFILE_OCID>

To list profiles, use the profile-summary list command:

oci optimizer profile-summary list --compartment-id, -c <COMPARTMENT_OCID>

Profile Descriptions

Some recommendations allow you to customize the recommendation profile to tailor the recommendations to your requirements. This section describes each profile.

Profiles for Downsize Underutilized Compute Instances

The following profiles are available for the Downsize underutilized compute instances recommendation type.

Average Methodology Profile Descriptions

If you choose the average methodology, the mean CPU utilization is used when evaluating compute instances. The profile options with the average methodology are:

  • Conservative (Average): The conservative profile identifies all compute instances that have had both:

    • an average CPU utilization less than 5% and
    • a maximum memory utilization less than 10%

    Supported VNICs are also considered. If a resize is recommended, the recommended core count has a projected average CPU utilization less than 10%.

  • Standard (Average): The standard profile identifies all compute instances that have had both:

    • an average CPU utilization less than 10% and
    • a maximum memory utilization less than 10%

    Supported VNICs are also considered. If a resize is recommended, the recommended core count has a projected average CPU utilization less than 20%.

  • Aggressive (Average): The Aggressive profile identifies all compute instances that have had both:

    • an average CPU utilization less than 15% and
    • a maximum memory utilization less than 10%.

    Supported VNICs are also considered. If a resize is recommended, the recommended core count has a projected average CPU utilization less than 30%.

P95 Methodology Profile Descriptions

If you choose the P95 methodology, a p-value of 95 is used to evaluate the CPU utilization threshold. The profile options with the P95 methodology are:

  • Conservative (P95): The conservative profile identifies all compute instances that have had both:

    • a P95 CPU utilization less than 5% and
    • a maximum memory utilization less than 10%

    Supported VNICs are also considered. If a resize is recommended, the recommended core count has a projected P95 CPU utilization less than 10%.

  • Standard (P95): The standard profile identifies all compute instances that have had both:

    • a P95 CPU utilization less than 10% and
    • a maximum memory utilization less than 10%

    Network throughput and supported VNICs are also considered. If a resize is recommended, the recommended core count has a projected P95 CPU utilization less than 20%.

  • Aggressive (P95): The aggressive profile identifies all compute instances that have had both:

    • a P95 CPU utilization of less than 15% and
    • a maximum memory utilization of less than 10%

    Network throughput and supported VNICs are also considered. If a resize is recommended, the recommended core count has a projected P95 CPU utilization less than 30%.

Profiles for Downsize Underutilized Load Balancers

The following profiles are available for the Downsize underutilized load balancers recommendation type. All profile options use the average methodology.

Profile Descriptions

For these profiles, the PeakBandwidth metric is used. Peak bandwidth measures the maximum bandwidth per second used during the specified interval. For more details, see Load Balancing Metrics. Cloud Advisor pulls the metric every three hours and averages the values over the specified evaluation period.

  • Conservative: The conservative profile identifies all load balancers for which the average of the maximum values for peak bandwidth usage over the evaluation period is less than 85% of the minimum network bandwidth.

  • Standard: The standard profile identifies all load balancers for which the average of the maximum values for peak bandwidth usage over the evaluation period is less than 90% of the minimum network bandwidth.

  • Aggressive: The aggressive profile identifies all load balancers for which the average of the maximum values for peak bandwidth usage over the evaluation period is less than 95% of the minimum network bandwidth.

Profiles for Rightsize Compute Instances

The following profiles are available for the Rightsize compute instances recommendation type.

Average Methodology Profile Descriptions

If you choose the average methodology, the mean CPU utilization is used when evaluating compute instances. The profile options with the average methodology are:

  • Conservative (Average): The conservative profile identifies all compute instances that have had both:

    • an average CPU utilization greater than 95% and
    • a maximum memory utilization greater than 95%
  • Standard (Average): The standard profile identifies all compute instances that have had both:

    • an average CPU utilization greater than 80% and
    • a maximum memory utilization greater than 80%
  • Aggressive (Average): The Aggressive profile identifies all compute instances that have had both:

    • an average CPU utilization greater than 60% and
    • a maximum memory utilization greater than 60%
P95 Methodology Profile Descriptions

If you choose the P95 methodology, a p-value of 95 is used to evaluate the CPU utilization threshold. The profile options with the P95 methodology are:

  • Conservative (P95): The conservative profile identifies all compute instances that have had both:

    • a P95 CPU utilization greater than 95% and
    • a maximum memory utilization greater than 95%
  • Standard (P95): The standard profile identifies all compute instances that have had both:

    • a P95 CPU utilization greater than 80% and
    • a maximum memory utilization greater than 80%
  • Aggressive (P95): The aggressive profile identifies all compute instances that have had both:

    • a P95 CPU utilization greater than 60% and
    • a maximum memory utilization greater than 60%

Profiles for Rightsize Load Balancers

The following profiles are available for the Rightsize load balancers recommendation type. All profile options use the average methodology.

Profile Descriptions

For these profiles, the PeakBandwidth metric is used. Peak bandwidth measures the maximum bandwidth per second used during the specified interval. For more details, see Load Balancing Metrics. Cloud Advisor pulls the metric every three hours and averages the values over the specified evaluation period.

  • Conservative: The conservative profile identifies all load balancers for which the average of the maximum values for peak bandwidth usage over the evaluation period is more than 95% of the maximum network bandwidth.

  • Standard: The standard profile identifies all load balancers for which the average of the maximum values for peak bandwidth usage over the evaluation period is more than 85% of the maximum network bandwidth.

  • Aggressive: The aggressive profile identifies all load balancers for which the average of the maximum values for peak bandwidth usage over the evaluation period is more than 75% of the maximum network bandwidth.