But in many cases, you will want to run several worker processes at the same time.

### Multiple Processes - Workers

For example (depending on your setup) you could have some tool like a Prometheus exporter in the same container that should have access to **each of the requests** that come.

* Keep commits discrete. Avoid including multiple unrelated changes in a single commit.
* Keep commits self-contained. Avoid spreading a single change across multiple commits. A single commit should make sense in isolation.

### Testing changes

After making changes, you can test your code in 2 ways:

1. Run tests.

All source code in Gradle is written to target one or more of these runtimes.
Most source code targets the daemon and the remaining code either targets a single runtime, for example the Gradle client, or is shared across multiple runtimes.

## Composition by architecture modules

Each [architecture module and platform](platforms.md) can contribute code to any of the runtimes.
Not every module contributes to every runtime.

    id("${model.projectId}_FlakyQuarantine_${os.asName()}_${arch.asName()}")
    name = "Flaky Test Quarantine - ${os.asName()} ${arch.asName()}"
    description = "Run all flaky tests skipped multiple times"

    applyDefaultSettings(os = os, arch = arch, buildJvm = BuildToolBuildJvm, timeout = 180)

    val testsWithOs = model.stages.filter {
        it.stageName in listOf(

The webhooks that you define will end up in the **OpenAPI** schema and the automatic **docs UI**.

!!! info
    The `app.webhooks` object is actually just an `APIRouter`, the same type you would use when structuring your app with multiple files.

* This also allows using automatic **client code generation** in many languages.

### Automatic docs

Interactive API documentation and exploration web user interfaces. As the framework is based on OpenAPI, there are multiple options, 2 included by default.

                firSymbolBuilder.typeBuilder.buildKtType(coneKotlinType)
            },
            firDiagnostic as KtPsiDiagnostic,
            token,
        )
    }
    add(FirJsErrors.WRONG_MULTIPLE_INHERITANCE) { firDiagnostic ->
        WrongMultipleInheritanceImpl(
            firSymbolBuilder.callableBuilder.buildCallableSymbol(firDiagnostic.a),
            firDiagnostic as KtPsiDiagnostic,
            token,

build:win_clang --host_platform=//tensorflow/tools/toolchains/win:x64_windows-clang-cl
build:win_clang --compiler=clang-cl
build:win_clang --linkopt=/FORCE:MULTIPLE
build:win_clang --host_linkopt=/FORCE:MULTIPLE
test:win_clang --linkopt=/FORCE:MULTIPLE
test:win_clang --host_linkopt=/FORCE:MULTIPLE

# Same config as above but for XLA, which has different toolchain paths
build:win_clang_xla --copt=/clang:-Weverything

  private val tasksExecutor = Executors.newCachedThreadPool(threadFactory("TaskFaker"))

  /**
   * True if this task faker has ever had multiple tasks scheduled to run concurrently. Guarded by
   * [TaskRunner.lock].
   */
  var isParallel = false

  /** Number of calls to [TaskRunner.Backend.execute]. Guarded by [TaskRunner.lock]. */
  var executeCallCount = 0