Rediger

Del via

Flaky tests sample report

Azure DevOps Services | Azure DevOps Server | Azure DevOps Server 2022

You can create a report, as shown in the following image, that lists the flaky tests occurring in pipeline runs that include test tasks. A flaky test is a test that provides different outcomes, such as pass or fail, even when there are no changes in the source code or execution environment. For more information about enabling flaky tests, see Manage flaky tests. For information on adding tests to a pipeline, see the Test task resources section later in this article.

Screenshot of Flaky Tests Table report.

Use the queries provided in this article to generate the following reports:

  • Flaky tests for build workflow
  • Flaky tests for release workflow
  • Flaky tests for a particular branch
  • Flaky tests for a particular test file
  • Flaky tests for a particular test owner.

Note

The sample queries in this article require v3.0-preview or later of the Analytics OData feed. Share your feedback.

Category Requirements
Access levels - Project member.
- At least Basic access.
Permissions By default, project members have permission to query Analytics and create views. For more information about other prerequisites regarding service and feature enablement and general data tracking activities, see Permissions and prerequisites to access Analytics.

Note

This article assumes you read Overview of sample reports using OData queries and have a basic understanding of Power BI.

Sample queries

You can use the following queries of the TestResultsDaily entity set to create different but similar pipeline flaky test reports. The TestResultsDaily entity set provides a daily snapshot aggregate of TestResult executions, grouped by test.

Note

To find available properties for filtering or reporting, see the Metadata reference for Test Plans Analytics and Metadata reference for Azure Pipelines. You can filter queries or return properties by using any Property value defined under an EntityType, or any NavigationPropertyBinding Path value listed for an EntitySet. Each EntitySet maps to an EntityType, which documents the data type for each property.

Flaky tests for a Build workflow

Use the following queries to view the flaky tests for a Build workflow pipeline.

Copy and paste the following Power BI query directly into the Get Data > Blank Query window. For more information, see Overview of sample reports using OData queries.

let
   Source = OData.Feed ("https://analytics.dev.azure.com/{organization}/{project}/_odata/v4.0-preview/TestResultsDaily?
$apply=filter("
                &"Pipeline/PipelineName eq '{pipelineName} "
                &"And Date/Date ge {startdate} "
                &"And Workflow eq 'Build') "
        &"/groupby((TestSK, Test/TestName), "
            &"aggregate( "
                &"ResultCount with sum as TotalCount, "
                &"ResultPassCount with sum as PassedCount, "
                &"ResultFailCount with sum as FailedCount, "
                &"ResultNotExecutedCount with sum as NotExecutedCount, "
                &"ResultNotImpactedCount with sum as NotImpactedCount, "
                &"ResultFlakyCount with sum as FlakyCount)) "
    &"/filter(FlakyCount gt 0) "
    &"/compute( "
    &"(FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlaykRate) "
    ,null, [Implementation="2.0",OmitValues = ODataOmitValues.Nulls,ODataVersion = 4]) 
in
    Source

Flaky tests for Release workflow

Use the following queries to view the flaky tests for a Release workflow pipeline.

Copy and paste the following Power BI query directly into the Get Data > Blank Query window. For more information, see Overview of sample reports using OData queries.

let
   Source = OData.Feed ("https://analytics.dev.azure.com/{organization}/{project}/_odata/v4.0-preview/TestResultsDaily?
$apply=filter("
                &"Pipeline/PipelineName eq '{pipelineName}' "
                &"And Date/Date ge {startdate}) "
        &"/groupby((TestSK, Test/TestName, Workflow), "
        &"aggregate( "
                &"ResultCount with sum as TotalCount, "
                &"ResultPassCount with sum as PassedCount, "
                &"ResultFailCount with sum as FailedCount, "
                &"ResultNotExecutedCount with sum as NotExecutedCount, "
                &"ResultNotImpactedCount with sum as NotImpactedCount, "
                &"ResultFlakyCount with sum as FlakyCount)) "
        &"/filter(FlakyCount gt 0) "
    &"/compute( "
    &"(FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlakyRate) "
    ,null, [Implementation="2.0",OmitValues = ODataOmitValues.Nulls,ODataVersion = 4]) 
in
    Source

Flaky tests filtered by branch

To view the failed tests of a pipeline for a particular branch, use the following queries. To create the report, carry out the following extra steps along with what is specified later in this article.

  • Expand Branch into Branch.BranchName.
  • Select Power BI Visualization Slicer and add the field Branch.BranchName to the slicer's Field.
  • Select the branch name from the slicer for which you need to see the outcome summary.

For more information about using slicers, see Slicers in Power BI.

Copy and paste the following Power BI query directly into the Get Data > Blank Query window. For more information, see Overview of sample reports using OData queries.

let
   Source = OData.Feed ("https://analytics.dev.azure.com/{organization}/{project}/_odata/v4.0-preview/TestResultsDaily?
$apply=filter("
                &"Pipeline/PipelineName eq '{pipelineName}' "
                &"And Date/Date ge {startdate} "
                &"And Workflow eq 'Build') "
        &"/groupby((TestSK, Test/TestName, Branch/BranchName), "
            &"aggregate( "
                &"ResultCount with sum as TotalCount, "
                &"ResultPassCount with sum as PassedCount, "
                &"ResultFailCount with sum as FailedCount, "
                &"ResultNotExecutedCount with sum as NotExecutedCount, "
                &"ResultNotImpactedCount with sum as NotImpactedCount, "
                &"ResultFlakyCount with sum as FlakyCount)) "
    &"/filter(FlakyCount gt 0) "
    &"/compute( "
    &"(FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlakyRate) "
    ,null, [Implementation="2.0",OmitValues = ODataOmitValues.Nulls,ODataVersion = 4]) 
in
    Source

Flaky tests filtered by test file

To view the flaky tests for a pipeline and a particular test file, use the following queries. To create the report, carry out the following extra steps along with what is defined later in this article.

  • Expand Test into Test.ContainerName.
  • Select Power BI Visualization Slicer and add the field Test.ContainerName to the slicer's Field.
  • Select the container name from the slicer for which you need to see the outcome summary.

Copy and paste the following Power BI query directly into the Get Data > Blank Query window. For more information, see Overview of sample reports using OData queries.

let
   Source = OData.Feed ("https://analytics.dev.azure.com/{organization}/{project}/_odata/v4.0-preview/TestResultsDaily?
$apply=filter("
                &"Pipeline/PipelineName eq '{pipelineName}' "
                &"And Date/Date ge {startdate}) "
        &"/groupby((TestSK, Test/TestName, Test/ContainerName), "
        &"aggregate( "
                &"ResultCount with sum as TotalCount, "
                &"ResultPassCount with sum as PassedCount, "
                &"ResultFailCount with sum as FailedCount, "
                &"ResultNotExecutedCount with sum as NotExecutedCount, "
                &"ResultNotImpactedCount with sum as NotImpactedCount, "
                &"ResultFlakyCount with sum as FlakyCount)) "
        &"/filter(FlakyCount gt 0) "
    &"/compute( "
    &"(FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlakyRate) "
    ,null, [Implementation="2.0",OmitValues = ODataOmitValues.Nulls,ODataVersion = 4]) 
in
    Source

Flaky tests filtered by test owner

To view the flaky tests for a pipeline and tests owned by a particular test owner, use the following queries. To create the report, carry out the following extra steps along with what is defined later in this article.

  • Expand Test into Test.TestOwner.
  • Select Power BI Visualization Slicer and add the field Test.TestOwner to the slicer's Field.
  • Select the test owner from the slicer for which you need to see the outcome summary.

Copy and paste the following Power BI query directly into the Get Data > Blank Query window. For more information, see Overview of sample reports using OData queries.

let
   Source = OData.Feed ("https://analytics.dev.azure.com/{organization}/{project}/_odata/v4.0-preview/TestResultsDaily?
$apply=filter("
                &"Pipeline/PipelineName eq '{pipelineName}' "
                &"And Date/Date ge {startdate}) "
        &"/groupby((TestSK, Test/TestName, Test/TestOwner), "
        &"aggregate( "
                &"ResultCount with sum as TotalCount, "
                &"ResultPassCount with sum as PassedCount, "
                &"ResultFailCount with sum as FailedCount, "
                &"ResultNotExecutedCount with sum as NotExecutedCount, "
                &"ResultNotImpactedCount with sum as NotImpactedCount, "
                &"ResultFlakyCount with sum as FlakyCount)) "
        &"/filter(FlakyCount gt 0) "
    &"/compute( "
    &"(FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlakyRate) "
    ,null, [Implementation="2.0",OmitValues = ODataOmitValues.Nulls,ODataVersion = 4]) 
in
    Source

Substitution strings and query breakdown

Replace the following strings with your values. Don't include the braces {} in your substitution. For example, if your organization name is "Fabrikam", replace {organization} with Fabrikam, not {Fabrikam}.

 

  • {organization} - Your organization name.
  • {project} - Your team project name.
  • {pipelinename} - Your pipeline name. Example: Fabrikam hourly build pipeline
  • {startdate} - The date to start your report. Format: YYYY-MM-DDZ. Example: 2021-09-01Z represents September 1, 2021. Don't enclose in quotes or brackets and use two digits for both, month and date.

Query breakdown

The following table describes each part of the query.

Query part

Description

$apply=filter(

Start filter() clause.

Pipeline/PipelineName eq '{pipelineName}'

Return test runs for the specified pipeline.

and CompletedOn/Date ge {startdate}

Return test runs on or after the specified date.

and Workflow eq 'Build'

Return test runs for a Build workflow.

)

Close filter() clause.

/groupby(

Start groupby() clause.

(TestSK, Test/TestName),

Group by the test name.

aggregate(

Start aggregate clause to calculate the test runs matching the filter criteria.

ResultCount with sum as TotalCount,

Count the total number of test runs as TotalCount.

ResultPassCount with sum as PassedCount,

Count the total number of passed test runs as PassedCount.

ResultFailCount with sum as FailedCount,

Count the total number of failed test runs as FailedCount.

ResultNotExecutedCount with sum as NotExecutedCount

Count the total number of not executed test runs as NotExecutedCount.

ResultNotImpactedCount with sum as NotImpactedCount,

Count the total number of not affected test runs as NotImpactedCount.

ResultFlakyCount with sum as FlakyCount

Count the total number of flaky test runs as FlakyCount.

))

Close aggregate() and groupby() clauses.

/filter(FlakyCount gt 0)

Filter out only those tests that were flaky at least once.

/compute(

Start compute() clause.

(FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlakyRate

For all the flaky tests, calculate the FlakyRate.

)

Close compute() clause.

(Optional) Rename query

You can rename the default query label, Query1, to something more meaningful. Enter a new name from the Query Settings pane.

Screenshot of Power BI query menu options, rename query.

Expand the Test column in Power BI

Expand the Test column to show the expanded entity Test.TestName. Expanding the column flattens the record into specific fields. To learn how, see Transform Analytics data to generate Power BI reports, Expand columns.

Change column data type

  1. In Power Query Editor, select the TotalCount, PassedCount, FailedCount, NotExecutedCount, NotImpactedCount, and FlakyCount columns. Select Data Type from the Transform menu, and then choose Whole Number.

  2. Select the FlakyRate column. Select Data Type from the Transform menu, and then choose Decimal Number.

For more information about changing the data type, see Transform Analytics data to generate Power BI reports, Transform a column data type.

Close the query and apply your changes

When you finish all your data transformations, select Close & Apply from the Home menu. This action saves the query and returns you to the Report tab in Power BI.

Screenshot of Power Query Editor Close and Apply option.

Create the Table report

  1. In Power BI, under Visualizations, select Table. Drag and drop the fields onto the Columns area.

    Screenshot of visualization fields selections for Flaky tests table report.

  2. Add the following fields to the Columns section in the order listed.

    • Test.TestName
    • TotalCount
    • PassedCount
    • FailedCount
    • FlakyCount
    • FlakyRate

Your report should look similar to the following image.

Screenshot of Sample Flaky Tests Table report.

Test task resources

  • Last updated on