Native Recommendations

CloudSpend Native Recommendations brings optimization findings generated by AWS, Azure, and Google Cloud into CloudSpend, giving you a single place to review provider generated savings opportunities, understand their impact, and prioritize actions.

What are Native Recommendations?

AWS, Azure, and Google Cloud each provide their own optimization services that analyze your cloud environment and generate recommendations. These include AWS Cost Optimization Hub, Azure Advisor, and Google Cloud Recommender.

Normally, these recommendations are available only within the respective cloud provider consoles. Native Recommendations brings these provider-generated findings into CloudSpend and presents them in a consistent structure.

CloudSpend does not independently derive or second-guess the provider's analysis. The recommendations, estimated savings, and supporting details come from the respective cloud provider. CloudSpend provides a centralized view, consistent presentation, spend context, and handling of overlapping recommendations where applicable.

Native Recommendations works alongside CloudSpend's own recommendations. You can switch between CloudSpend and Native Recommendations from the Recommendations page.

Benefits of Native Recommendations

Native Recommendations help you:

• Review recommendations from multiple cloud providers in one place: View findings from AWS Cost Optimization Hub, Azure Advisor, and Google Cloud Recommender in a consistent format.
• Understand savings in the context of actual spend: Compare potential monthly savings with the account's spend over the preceding 30 days to understand the relative impact.
• Prioritize optimization opportunities: Review recommendations by source, recommendation type, severity, and estimated savings to identify opportunities that need attention.
• Evaluate proposed changes: View the resource's current configuration alongside the provider's preferred configuration before taking action.
• Receive consolidated savings estimates: CloudSpend groups overlapping recommendations where multiple provider recommendations represent the same underlying decision, preventing mutually exclusive options from inflating the total savings.
• Leverage existing integration resources: Native Recommendations uses the existing S3 bucket, BigQuery dataset, and Azure application registration configured for the cost account.
• Share recommendations with stakeholders: Schedule Native Recommendations reports and deliver summary or detailed findings by email.

CloudSpend Recommendations vs. Native Recommendations

CloudSpend provides two complementary sets of recommendations. They use different sources of information and can therefore identify different optimization opportunities.

Neither recommendation set replaces the other. Native Recommendations provides the cloud provider's perspective, while CloudSpend Recommendations provides CloudSpend's own analysis. Reviewing both can help you identify a broader range of optimization opportunities.

Native Recommendations use cases

Review optimization opportunities across multiple clouds

If you manage AWS, Azure, and Google Cloud environments, you can review recommendations from all three providers within CloudSpend instead of moving between multiple provider consoles.

Prioritize recommendations based on actual spend

A recommendation's estimated savings can have different significance depending on the account's overall spend. Native Recommendations shows potential savings alongside the preceding 30 days of spend and expresses the savings as a percentage of that spend.

This helps you identify opportunities that have a meaningful impact on your cloud costs.

Evaluate a recommended resource change

When a provider recommends changing a resource configuration, open the recommendation details to compare the current configuration with the recommended configuration.

This lets you understand the proposed change before deciding whether to act on it.

Evaluate different commitment options

Cloud providers can present multiple recommendations for the same underlying purchasing decision. For example, Azure can provide alternatives based on commitment terms and utilization look-back periods, while Google Cloud can provide alternatives based on commitment terms and algorithms.

CloudSpend groups these overlapping alternatives and selects the option with the highest estimated savings as the primary recommendation. The other alternatives remain available on the detail page for comparison.

Share recommendations with stakeholders

FinOps and cloud teams can schedule Native Recommendation reports and share them with stakeholders who do not regularly access CloudSpend.

Summary reports provide an overview of the recommendations, while Investigate reports provide detailed resource-level information.

Combine provider recommendations with CloudSpend analysis

You can switch between CloudSpend and Native recommendations from the Recommendations page. Reviewing both helps you compare provider generated findings with CloudSpend's own analysis of resource configuration and performance.

Prerequisites

Native Recommendations reads optimization data generated by the cloud provider. The required setup varies by cloud provider.

Native Recommendations is configured in addition to the existing billing data integration for the cost account. CloudSpend reuses the existing resources used for billing data.

For AWS and GCP, the configuration is done once per cost account and covers everything beneath the account. For Azure, access is assigned at the subscription-level.

Configure AWS Recommendations

The configuration steps vary depending on your cloud provider. AWS delivers Cost Optimization Hub recommendations through a data export to Amazon S3.

In CloudSpend, while integrating the AWS account, configure the Cost Optimization Recommendations Export section.

If you have already created the required Cost Optimization Recommendations data export for the account, select Yes. If you have not, leave No selected and follow the steps below.

Create the recommendations export in AWS

1. Sign in to the AWS Management Console and select Billing and Cost Management.
2. In the left navigation pane, select Data Exports, then select Create.
3. On the Create export page, under Export type, select Standard data export.
4. Enter CostRecommendationsReports as the Export name. This name must match exactly for CloudSpend to recognize the export.
5. Under Data table content settings, select Cost optimization recommendations.
6. Under Data table configurations, leave Include all recommendations unchecked. This allows AWS to exclude incompatible recommendations and retain the applicable recommendation based on the configured export behavior.
7. Under Data export delivery options, select gzip as the Compression type and text/csv as the File format. For File versioning, select Overwrite existing data export file.
8. Under Data export storage settings, select the same Amazon S3 bucket used for your billing data export.
9. Enter recommendations as the S3 path prefix. This value must match exactly.
10. Click Create.

The IAM role used by CloudSpend for the account must also have the following permissions:

• cost-optimization-hub:ListRecommendations 
• cost-optimization-hub:GetRecommendation

Recommendations appear in CloudSpend after the first export file is delivered to the configured S3 bucket.

Note: The export name CostRecommendationsReports and the S3 path prefix recommendations must match exactly. If either value is incorrect, AWS can create the export successfully, but CloudSpend will not recognize the recommendation data.

Configure Azure Recommendations

Azure does not require a separate data export. CloudSpend reads Azure Advisor recommendations directly from the subscription through the Advisor API. In CloudSpend, while integrating the Azure account, configure the Azure Advisor Access section.

If you have already assigned the Reader role to the application registered with CloudSpend, select Yes. If you have not assigned the role, leave No selected and follow the steps below.

Assign the Reader role

1. Sign in to the Microsoft Azure portal and go to Subscriptions.
2. Select the Azure subscription you are integrating with CloudSpend.
3. Select Access control (IAM) and select Add > Add role assignment.
4. Select the built-in Reader role.
5. Assign the role to the application registered with CloudSpend and select Review + assign.

Recommendations appear automatically after the role is assigned. Because Azure permissions are granted per subscription, different subscriptions under the same cost account can have different configuration states. The subscription picker remains available so you can identify which subscriptions are configured.

Configure GCP Recommendations

Google Cloud provides Recommender data through a BigQuery data transfer. In CloudSpend, while integrating the GCP account, configure the Recommender Data Transfer section.

If you have already set up the Recommender data transfer in BigQuery, select Yes. If you have not set up the transfer, leave No selected and follow the steps below.

Create the Recommender data transfer

1. In the Google Cloud console, go to BigQuery > Data transfers and select Createtransfer.
2. Select Recommender V1 as the Source.
3. Under Schedule options, set the Repeat frequency to 24 hours.
4. Under Destination settings, select the same dataset used for your billing data export.
5. Under Data source details, enter the Organization ID associated with the GCP account integrated with CloudSpend. Use the same Organization ID configured for your billing export.
6. Click Save.

The transfer exports Recommender data to BigQuery, and CloudSpend reads the recommendation data from the same dataset used for the billing export. Recommendations appear automatically after the first transfer run completes.

View Native Recommendations

After the required configuration is complete and recommendation data is available, you can view Native Recommendations from Reports > Recommendations. Select Native to view recommendations generated by your cloud provider.

The Native Recommendations page provides an overview of the optimization opportunities available for the selected cloud account.

The numbers in the above screenshot correspond to the UI elements described below.

1. Potential Savings: Displays the total potential monthly savings available from the recommendations for the selected account and the number of recommendations that contribute to the savings.
2. Savings by Recommendation Type: Displays potential savings grouped by recommendation type. You can view the savings as a pie chart, bar chart, or table.
3. Savings (%): Shows the account's spend over the preceding 30 days and the percentage of that spend that can potentially be saved through the available recommendations.
4. Recommendations: Lists the recommendation sources available for the selected account. For each source, you can view when the recommendations were last updated, the number of recommendations, estimated monthly savings, and the severity distribution
5. Investigate: Click Investigate for a recommendation source to view the resources affected by the recommendations.
6. ScheduleReport: Schedule a Native Recommendations report for the selected account in PDF format.
7. Share: Share the recommendations with other users as PDF or email.

Investigate recommendations

Select Investigate for a recommendation source to open the Investigate page. This page provides a resource level view of the recommendations for the selected source.

The page displays the following information:

• Affected Resources: Shows the number of resources flagged by the selected recommendation source.
• Estimated Monthly Savings: Displays the estimated monthly savings across the affected resources.
• Severity: Displays the highest severity among the affected recommendations.

You can use the Search field to find a specific resource and the severity filter to filter recommendations by severity.

The recommendation table displays the following details:

• Resource ID: Identifies the resource affected by the recommendation. Select a resource ID to view the complete recommendation details.
  Type: Displays the type of recommendation, such as Reservations, Savings Plan, or Usage Optimization.
• Resource Type: Displays the type of cloud resource associated with the recommendation.
• Region: Displays the region or location associated with the affected resource.
• Savings ($): Displays the estimated savings for the individual recommendation.

Recommendation details

The Recommendation Details page displays the recommendation data provided by the cloud provider. It includes the provider's supporting data and a comparison between the resource's current configuration and the configuration recommended by the provider. 
Click the required recommendation on the Investigate page to open the Recommendation Details page.

The information displayed on this page varies depending on the cloud provider and recommendation type. CloudSpend displays the data published by the respective cloud provider, so the fields and their structure can differ across AWS, Azure, and GCP.

The details can include information such as:

• Resource type and resource ID
• Region or location
• Recommendation type or subcategory
• Savings amount and currency
• Quantity and resource count
• Recommended resource configuration
• Commitment term or look back period, where applicable
• Provider specific identifiers and supporting information

For recommendations with multiple alternatives, such as different commitment terms or look back periods, use the available selectors to review the applicable options.

Understand the recommendation details

The following terms are used throughout Native Recommendations.

• Source: The cloud provider advisor or service that generated the recommendation. For AWS, sources can include areas such as Billing and Cost Management and Compute Optimizer. For Azure, sources follow Advisor categories such as Cost, Security, High Availability, and Operational Excellence.
• Severity: The provider's assessment of the recommendation's importance, normalized in CloudSpend as High, Medium, Low, or Very Low.
• Recommendation type: The type of optimization being proposed. Examples include purchasing savings plans, purchasing reserved instances, rightsizing, upgrading, migrating, or deleting an unused resource.

Recommendation freshness

Native Recommendations are refreshed automatically based on the recommendation data received from the cloud provider.

Each recommendation source displays an Updated N hours ago timestamp. This indicates how long ago the recommendation data was last synced from the cloud provider to CloudSpend. You do not need to manually refresh the recommendations.

To find out how recently a specific recommendation was generated by the cloud provider, open the recommendation's Details page. The recommendation details include the provider's recommendation generation timestamp, enabling you to distinguish between when the recommendation was generated and when it was last synced to CloudSpend.

Once the integration is configured, new recommendation data is fetched and made available in CloudSpend as part of the regular synchronization process.

How CloudSpend handles overlapping recommendations

Cloud providers can present multiple recommendations for the same underlying purchasing decision. These recommendations may represent different commitment terms, utilization look back periods, or calculation methods.

For example, Azure Advisor can provide the same reservation or savings plan recommendation across different commitment terms and utilization look back periods. Google Cloud can provide multiple committed use discount recommendations for the same commitment across different terms and algorithms.

AWS Cost Optimization Hub already consolidates recommendations generated by AWS services such as Compute Optimizer and Trusted Advisor. CloudSpend brings these consolidated Cost Optimization Hub recommendations into Native Recommendations.

For Azure and GCP, some of the recommendations provided by the cloud provider are alternatives rather than additional savings opportunities. A user would act on one option, not all of them.

What CloudSpend does

CloudSpend handles overlapping recommendations for Azure and GCP by grouping alternatives that represent the same underlying decision and retaining the option with the highest estimated savings as the primary recommendation.

For AWS, Cost Optimization Hub itself serves as the aggregation point for recommendations from AWS services such as Compute Optimizer and Trusted Advisor. CloudSpend uses the recommendations provided by Cost Optimization Hub rather than separately combining recommendations from these underlying services.

For Azure, if alternatives have the same estimated savings, CloudSpend uses the shorter look back period and then the more recently updated recommendation as tie breakers.

For GCP, the highest savings value determines the preferred commitment term. When the algorithms produce a genuine tie, CloudSpend applies a tie breaker to select the primary recommendation.

The alternatives that were not selected for Azure and GCP are retained and displayed on the recommendation detail page. You can use the selector to review and compare the available options.

This deduplication applies only where a provider genuinely presents overlapping alternatives. It does not merge recommendations that represent independent actions.

Because of this, the total savings shown in CloudSpend can be lower than the total shown in the Azure or GCP console. CloudSpend avoids adding mutually exclusive options together, so the displayed savings reflect the opportunity that can actually be acted upon.

Schedule Native Recommendation reports

Native Recommendations can be delivered by email on a schedule, enabling you to share recommendations with stakeholders who do not regularly access CloudSpend.

When scheduling a report, you can select:

• Source: Select CloudSpend or Native.
• Type: For Native Recommendations, select Summary or Investigate.
• Format:
  • Summary reports are available in PDF format.
  • Investigate reports are available in PDF or CSV format.
  • Investigate reports generate one file for each recommendation source.

Only accounts whose integration is currently syncing successfully can be scheduled.

You can also schedule a report directly from the Recommendations or Investigate pages. When you do this, the source, type, and account are prefilled based on the page you are viewing.

Related topics

• Recommendations Report
• AWS Recommendations Report
• Azure Recommendations Report
• GCP Recommendations Report
• Insights