Kopīgot, izmantojot

Troubleshooting in Azure carbon optimization (Preview)

  • Raksts

This article helps you with the diagnosis and troubleshooting of some of the common issues you might encounter when using Azure carbon optimization (Preview).

1. Access issues

1.1. Access denied on Carbon Optimization

Problem: Permissions issues when accessing the Carbon Optimization blade.

Solution:

  • Request appropriate permissions.
  • Carbon emissions data is accessible to Carbon Optimization Reader, Subscription Owner, and Subscription Contributor roles. This role allows you to view carbon emissions data on the Emission Trends and Emission Details pages.
  • Permissions are set at the subscription level.

Steps to Assign Carbon Optimization Reader Role:

To assign the Carbon Optimization Reader role, you must be a subscription owner or contributor.

  1. Navigate to the Subscriptions page in the Azure portal.
  2. Select the subscriptions you want to assign the role to.
  3. Select Access control (IAM).
  4. Select Add role assignment.
  5. Select Carbon Optimization Reader from the Role drop-down list.
  6. Select the user or group you want to assign the role to.
  7. Select Save.

2. Data limitations

2.1. Viewing data for more than 100 subscriptions

Problem: Data is available for a maximum of 100 subscriptions selected at a time when viewing carbon emissions data within the carbon optimization blade.

Current behavior: If you select more than 100 subscriptions, data is only displayed for the first 100 subscriptions selected.

Solution: Ensure you select no more than 100 subscriptions to view accurate and complete data.

Tips for selecting relevant subscriptions:

  • Review your selection to ensure it doesn't exceed 100 subscriptions. The subscription selection filter displays the number of subscriptions selected.
  • Select subscriptions most relevant to your carbon optimization goals (for example, highest usage or emissions).
  • Perform analysis in multiple steps, selecting different sets of up to 100 subscriptions at the time.

Note

To access data for more than 100 subscriptions or any other large scale data operations, use the Microsoft Azure emissions insights capability, available as part of the Sustainability data solutions in Fabric.

2.2. Maximum 5,000 results returned in an API query

Problem: For any API query, the result set returns a maximum of 5,000 items, even with pagination. 

Solution: Limit the input parameters in your query.

Helpful tips:

  • Query data for a single subscription at a time.
  • Query for a single or a subset of resource groups.
  • Query data for one month at a time.

2.3. Twelve month lookback period for emissions data

Problem: Azure Carbon Optimization offers emissions data for your resources for the last twelve months.

Solution: For data beyond the last twelve months, consider using:

3. Emissions data issues

All or most Scope 2 emissions are 0

Explanation: It's expected as we show Scope 2 Market data and not Scope 2 Location data. For more information, see the calculation methodology.

4. API issues

4.1. API AAD authentication error

Error: Failed to send request to Microsoft.Carbon due to AAD authentication error

Explanation: This error is likely when your tenant admin has disabled the app on your end.

Solution:

  • Reach out to your tenant admin to check if the "Prod-AzureSustainability" is disabled in your tenant. If it's disabled, enable it.

4.2. One month date scope supported in certain API Report types

Problem: When attempting to request data spanning multiple months, an error occurs, suggesting a request should be made for one month at a time.

Example error:

"error": {
          "code": "InvalidRequestPropertyValue",
          "message": "The start date 11/01/2023 00:00:00 and end date 12/01/2023 00:00:00 should be equal for report type ItemDetailsReport."
        }

Expected behavior: 'ItemDetailsReport' and 'TopItemsSummaryReport' only support a single month at a time.

Solution:

  • Query data for a single month when using the 'ItemDetailsReport' and 'TopItemsSummaryReport' API report types.
  • You must enter the first day of each month to get emissions data for the whole month.

4.3. Request Throttling

Problem: There's a limit of 100 queries per minute (QPM) per tenant when calling the API.

Expected behavior: Exceeding this limit may result in throttling and other requests may be queued or rejected until the rate drops below the limit.

Solution: To avoid throttling, monitor and manage your query rate, especially during periods of high demand.

5. General limitations

5.1. Can’t display data for the selected subscriptions in the Carbon Optimization page

Problem: The error message Data can't be displayed for the selected subscriptions might be displayed for many reasons:

  • Missing Carbon Optimization Reader role for some subscriptions.
  • New subscriptions with no data available yet (data takes at least a month to appear).
  • Subscriptions with no resource usage, hence no emissions data.

Solution:

  • Ensure you have the necessary permissions. For more information, see Access Denied on Carbon Optimization.
  • Wait for data to become available for new subscriptions.
  • Check subscription usage in the Cost Analysis blade or Resource page.

5.2. One hour wait for tenant registration at first time visit

Problem: Up to one hour wait for some tenants on their first visit.

Expected behavior: If you are on a tenant that is viewing the Azure carbon optimization blade for the first time, there's a 30 minutes to one hour wait for your tenant to be registered in the system. During this time, you may see a message indicating that tenant registration is in progress, or a blank page may be shown. When the registration is complete, you can see your data. Once your tenant is registered, other users in your organization on the same tenant have access to the data without any wait time.

5.3. Data latency

Problem: Changes to resource configuration are reflected in the emissions data of the following month.

Solution:

  • Data is updated by the 19th day of the following month.
  • Plan your analysis accordingly, taking into account this data update schedule.