Skip to content

sample-test-analytics-flaky-tests.md

Latest commit

 

History

History
553 lines (464 loc) · 18.7 KB

sample-test-analytics-flaky-tests.md

File metadata and controls

553 lines (464 loc) · 18.7 KB
title Pipeline flaky test sample Power BI reports
titleSuffix Azure DevOps
description Learn how to generate a list of flaky tests Power BI report for a given pipeline in the project.
ms.subservice azure-devops-analytics
ms.reviewer desalg
ms.manager mijacobs
ms.author chcomley
ms.custom powerbisample, engagement-fy23
author chcomley
ms.topic sample
monikerRange <=azure-devops
ms.date 01/25/2023

Flaky tests sample report

[!INCLUDE version-lt-eq-azure-devops]

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.

:::image type="content" source="media/pipeline-test-reports/flaky-tests-table-report.png" alt-text="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.

[!INCLUDE temp]

[!INCLUDE prerequisites-simple]

[!INCLUDE temp]

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.

[!INCLUDE temp]

Flaky tests for a Build workflow

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

Power BI query

[!INCLUDE temp]

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

OData query

[!INCLUDE temp]

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 FlakyRate
)

Flaky tests for Release workflow

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

Power BI query

[!INCLUDE temp]

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

OData query

[!INCLUDE temp]

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)

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.

Power BI query

[!INCLUDE temp]

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

OData query

[!INCLUDE temp]

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)

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.

Power BI query

[!INCLUDE temp]

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

OData query

[!INCLUDE temp]

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)

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.

Power BI query

[!INCLUDE temp]

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

OData query

[!INCLUDE temp]

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)

Substitution strings and query breakdown

[!INCLUDE temp]  

  • {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.

:::row::: :::column span="1"::: Query part :::column-end::: :::column span="1"::: Description :::column-end::: :::row-end::: :::row::: :::column span="1"::: $apply=filter( :::column-end::: :::column span="1"::: Start filter() clause. :::column-end::: :::row-end::: :::row::: :::column span="1"::: Pipeline/PipelineName eq '{pipelineName}' :::column-end::: :::column span="1"::: Return test runs for the specified pipeline. :::column-end::: :::row-end::: :::row::: :::column span="1"::: and CompletedOn/Date ge {startdate} :::column-end::: :::column span="1"::: Return test runs on or after the specified date. :::column-end::: :::row-end::: :::row::: :::column span="1"::: and Workflow eq 'Build' :::column-end::: :::column span="1"::: Return test runs for a Build workflow. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ) :::column-end::: :::column span="1"::: Close filter() clause. :::column-end::: :::row-end::: :::row::: :::column span="1"::: /groupby( :::column-end::: :::column span="1"::: Start groupby() clause. :::column-end::: :::row-end::: :::row::: :::column span="1"::: (TestSK, Test/TestName), :::column-end::: :::column span="1"::: Group by the test name. :::column-end::: :::row-end::: :::row::: :::column span="1"::: aggregate( :::column-end::: :::column span="1"::: Start aggregate clause to calculate the test runs matching the filter criteria. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ResultCount with sum as TotalCount, :::column-end::: :::column span="1"::: Count the total number of test runs as TotalCount. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ResultPassCount with sum as PassedCount, :::column-end::: :::column span="1"::: Count the total number of passed test runs as PassedCount. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ResultFailCount with sum as FailedCount, :::column-end::: :::column span="1"::: Count the total number of failed test runs as FailedCount. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ResultNotExecutedCount with sum as NotExecutedCount :::column-end::: :::column span="1"::: Count the total number of not executed test runs as NotExecutedCount. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ResultNotImpactedCount with sum as NotImpactedCount, :::column-end::: :::column span="1"::: Count the total number of not affected test runs as NotImpactedCount. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ResultFlakyCount with sum as FlakyCount :::column-end::: :::column span="1"::: Count the total number of flaky test runs as FlakyCount. :::column-end::: :::row-end::: :::row::: :::column span="1"::: )) :::column-end::: :::column span="1"::: Close aggregate() and groupby() clauses. :::column-end::: :::row-end::: :::row::: :::column span="1"::: /filter(FlakyCount gt 0) :::column-end::: :::column span="1"::: Filter out only those tests that were flaky at least once. :::column-end::: :::row-end::: :::row::: :::column span="1"::: /compute( :::column-end::: :::column span="1"::: Start compute() clause. :::column-end::: :::row-end::: :::row::: :::column span="1"::: (FlakyCount div cast(TotalCount, Edm.Decimal)) mul 100 as FlakyRate :::column-end::: :::column span="1"::: For all the flaky tests, calculate the FlakyRate. :::column-end::: :::row-end::: :::row::: :::column span="1"::: ) :::column-end::: :::column span="1"::: Close compute() clause. :::column-end::: :::row-end:::

[!INCLUDE temp]

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. From the 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.

[!INCLUDE temp]

Create the Table report

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

    :::image type="content" source="media/pipeline-test-reports/visualizations-flaky-test-table.png" alt-text="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.

:::image type="content" source="media/pipeline-test-reports/flaky-tests-table-report.png" alt-text="Screenshot of Sample Flaky Tests Table report.":::

[!INCLUDE temp]

Related articles

[!INCLUDE temp]