How can I generate accurate code coverage reports using Coverlet in .NET projects?

Use Coverlet through the test workflow that matches the project and pipeline setup:

1. For standard dotnet test workflows, use the collector integration:

      dotnet test --collect:"XPlat Code Coverage"
   

   This produces a coverage.cobertura.xml file under the TestResults directory. This is a cross-platform option and fits well in build systems that rely on the .NET CLI.
2. If the build already uses MSBuild-based test execution, use the MSBuild integration:

      dotnet test /p:CollectCoverage=true /p:CoverletOutputFormat=cobertura
   

   This also produces a coverage.cobertura.xml file.
3. For Microsoft Testing Platform-based projects, add the coverlet.MTP package and run:

      dotnet add package coverlet.MTP
      dotnet test --coverlet
   

   This generates coverage.json in the current directory by default.

Recommended approach

• Prefer coverlet.collector with dotnet test --collect:"XPlat Code Coverage" for modern cross-platform .NET test workflows.
• Use Cobertura output when the report needs to be published in CI/CD systems.
• Use a report generator to turn collected coverage data into human-readable reports such as HTML.

Useful Coverlet options with Microsoft Testing Platform

• --coverlet-output-format <format> supports json, lcov, opencover, cobertura, and teamcity.
• --coverlet-include <filter> and --coverlet-exclude <filter> control which assemblies are measured.
• --coverlet-exclude-by-file <pattern> excludes files by glob pattern.
• --coverlet-exclude-by-attribute <attribute> excludes code marked with specific attributes.
• --coverlet-include-test-assembly includes the test assembly in the report.
• --coverlet-skip-auto-props skips auto-implemented properties.

Best practices for consistent results

• Standardize on one collection mode across projects in the same solution or pipeline.
• Generate Cobertura output when the pipeline needs to publish coverage results.
• Use include and exclude filters deliberately so coverage reflects the actual system under test instead of test infrastructure or generated code.
• Keep package versions current. For Microsoft Testing Platform compatibility, the documented combinations are:
  • Microsoft.Testing.Extensions.CodeCoverage 18.1.x with MTP 2.0.x
  • 18.0.x with 1.8.x
  • 17.14.x with 1.6.2 Using the latest compatible versions is recommended.

CI/CD guidance

In Azure Pipelines:

• On Windows, the built-in collector can be used with DotNetCoreCLI@2:

    - task: DotNetCoreCLI@2
      inputs:
        command: test
        projects: '**/*Tests/*.csproj'
        arguments: '--configuration $(buildConfiguration) --collect "Code Coverage"'
  

  This automatically publishes coverage data to the server, and the *.coverage file can be downloaded from the build summary for viewing in Visual Studio.
• For Linux or macOS, use Coverlet.
• To publish coverage results to the server, configure the coverage tool to generate Cobertura or JaCoCo format, then publish those results. When using coverlet.collector, avoid adding extra DataCollectionRunSettings arguments because the XPlat Code Coverage collector already produces a Cobertura report.

Common pitfalls to avoid

• Mixing different coverage collection approaches across projects without standardizing output format.
• Expecting the built-in .coverage binary output to be directly human-readable.
• Adding unnecessary extra collector settings when XPlat Code Coverage already emits Cobertura.
• Using incompatible Microsoft Testing Platform and code coverage package versions.

If a readable report is needed after collection, use a report generator on the Cobertura output. Coverlet collects the coverage data; report generation is a separate step.

• Use code coverage for unit testing
• Use code coverage for unit testing
• Code coverage
• Code coverage
• Build, test, and deploy .NET Core projects