### Anotações de Tipo de Retorno Inválido { #invalid-return-type-annotations }

Mas quando você retorna algum outro objeto arbitrário que não é um tipo Pydantic válido (por exemplo, um objeto de banco de dados) e você o anota dessa forma na função, o FastAPI tentará criar um modelo de resposta Pydantic a partir dessa anotação de tipo e falhará.

// scss-docs-end tooltip-css-vars\n\n  z-index: var(--#{$prefix}tooltip-zindex);\n  display: block;\n  margin: var(--#{$prefix}tooltip-margin);\n  @include deprecate(\"`$tooltip-margin`\", \"v5\", \"v5.x\", true);\n  // Our parent element can be arbitrary since tooltips are by default inserted as a sibling of their target element.\n  // So reset our font and text properties to avoid inheriting weird values.\n  @include reset-text();\n  @include font-size(var(--#{$prefix}tooltip-font-size));\n  //...

    /*
     * This code performs one System.nanoTime() more than necessary, and in return, the time to
     * execute Queue#drainTo is not added *on top* of waiting for the timeout (which could make
     * the timeout arbitrarily inaccurate, given a queue that is slow to drain).
     */
    long deadline = System.nanoTime() + unit.toNanos(timeout);
    int added = 0;
    while (added < numElements) {

   * f(x)^c with greater than expected probability. The test for funneling is merely a test for
   * 1-bit characteristics.
   *
   * <p>There is more general code provided by Bob Jenkins to test arbitrarily sized characteristics
   * using the magic of gaussian elimination: http://burtleburtle.net/bob/crypto/findingc.html.
   */
  static void checkNo2BitCharacteristics(HashFunction function) {
    Random rand = new Random(0);

    // Service b startup takes at least 353 millis, but starting the timer is delayed by at least
    // 150 milliseconds. so in a perfect world the timing would be 353-150=203ms, but since either
    // of our sleep calls can be arbitrarily delayed we should just assert that there is a time
    // recorded.
    assertThat(startupTimes.get(b)).isNotNull();
  }

  public void testServiceStartStop() {
    Service a = new NoOpService();

[CSI Ephemeral Volume](https://github.com/kubernetes/enhancements/tree/master/keps/sig-storage/596-csi-inline-volumes) feature allows CSI volumes to be specified directly in the pod specification for ephemeral use cases. They can be used to inject arbitrary states, such as configuration, secrets, identity, variables or similar information, directly inside pods using a mounted volume. This was initially introduced in 1.15 as an alpha feature, and it moved to GA. This feature is used by some CSI drivers...

### Anotaciones de Tipos de Retorno Inválidas { #invalid-return-type-annotations }

Pero cuando devuelves algún otro objeto arbitrario que no es un tipo válido de Pydantic (por ejemplo, un objeto de base de datos) y lo anotas así en la función, FastAPI intentará crear un modelo de response de Pydantic a partir de esa anotación de tipo, y fallará.

    /*
     * This code performs one System.nanoTime() more than necessary, and in return, the time to
     * execute Queue#drainTo is not added *on top* of waiting for the timeout (which could make
     * the timeout arbitrarily inaccurate, given a queue that is slow to drain).
     */
    long deadline = System.nanoTime() + unit.toNanos(timeout);
    int added = 0;
    while (added < numElements) {

   * f(x)^c with greater than expected probability. The test for funneling is merely a test for
   * 1-bit characteristics.
   *
   * <p>There is more general code provided by Bob Jenkins to test arbitrarily sized characteristics
   * using the magic of gaussian elimination: http://burtleburtle.net/bob/crypto/findingc.html.
   */
  static void checkNo2BitCharacteristics(HashFunction function) {
    Random rand = new Random(0);

Podrías establecer un **número arbitrario** para alcanzar, por ejemplo, algo **entre 50% a 90%** de utilización de recursos. El punto es que esas son probablemente las principales cosas que querrás medir y usar para ajustar tus implementaciones.