Note over dep: Run code up to yield
    opt raise
        dep -->> handler: Raise HTTPException
        handler -->> client: HTTP error response
        dep -->> dep: Raise other exception
    end
    dep ->> operation: Run dependency, e.g. DB session
    opt raise
        operation -->> handler: Raise HTTPException
        handler -->> client: HTTP error response
        operation -->> dep: Raise other exception

Можно добавить собственный обработчик исключений с помощью `@app.exception_handler()`:

```Python hl_lines="5-7  13-18  24"
{!../../../docs_src/handling_errors/tutorial003.py!}
```

Здесь, если запросить `/unicorns/yolo`, то *операция пути* вызовет `UnicornException`.

Но оно будет обработано `unicorn_exception_handler`.

Таким образом, вы получите чистую ошибку с кодом состояния HTTP `418` и содержимым JSON:

// This cannot be used concurrently.
func (m *muxClient) roundtrip(h HandlerID, req []byte) ([]byte, error) {
	if m.init {
		return nil, errors.New("mux client already used")
	}
	m.init = true
	m.singleResp = true
	msg := message{
		Op:         OpRequest,
		MuxID:      m.MuxID,
		Handler:    h,
		Flags:      m.BaseFlags | FlagEOF,
		Payload:    req,

```Python hl_lines="5-7  13-18  24"
{!../../../docs_src/handling_errors/tutorial003.py!}
```

Nesse cenário, se você fizer uma requisição para `/unicorns/yolo`, a *operação de caminho* vai lançar (`raise`) o `UnicornException`.

Essa exceção será manipulada, contudo, pelo `unicorn_exception_handler`.

		err = mg.bucketLoader(GlobalContext, metricValues, mg.cache, mg.buckets)
	} else {
		err = mg.loader(GlobalContext, metricValues, mg.cache)
	}

	// There is no way to handle errors here, so we panic the current goroutine
	// and the Metrics API handler returns a 500 HTTP status code. This should
	// normally not happen, and usually indicates a bug.
	logger.CriticalIf(GlobalContext, errors.Wrap(err, "failed to get metrics"))

				}
			}
		}

		return metrics
	})
	return mg
}

// metricsResourceHandler is the prometheus handler for resource metrics
func metricsResourceHandler() http.Handler {
	return metricsHTTPHandler(resourceCollector, "handler.MetricsResource")

  @BeforeEach fun setUp(server: MockWebServer?) {
    this.server = server
  }

  @AfterEach fun tearDown() {
    ProxySelector.setDefault(DEFAULT_PROXY_SELECTOR)
    CookieManager.setDefault(DEFAULT_COOKIE_HANDLER)
    ResponseCache.setDefault(DEFAULT_RESPONSE_CACHE)
  }

  @Test fun durationDefaults() {
    val client = clientTestRule.newClient()
    assertThat(client.callTimeoutMillis).isEqualTo(0)

そして、この例外をFastAPIでグローバルに処理したいと思います。

カスタム例外ハンドラを`@app.exception_handler()`で追加することができます:

```Python hl_lines="5 6 7  13 14 15 16 17 18  24"
{!../../../docs_src/handling_errors/tutorial003.py!}
```

ここで、`/unicorns/yolo`をリクエストすると、*path operation*は`UnicornException`を`raise`します。

しかし、これは`unicorn_exception_handler`で処理されます。

そのため、HTTPステータスコードが`418`で、JSONの内容が以下のような明確なエラーを受け取ることになります:

```JSON

        @classmethod
        def __get_pydantic_json_schema__(
            cls, core_schema: CoreSchema, handler: GetJsonSchemaHandler
        ) -> JsonSchemaValue:
            return {"type": "string", "format": "email"}

        @classmethod
        def __get_pydantic_core_schema__(
            cls, source: Type[Any], handler: Callable[[Any], CoreSchema]
        ) -> CoreSchema:

	"sort"
	"strings"
	"sync"

	"github.com/tinylib/msgp/msgp"
)

// Recycler will override the internal reuse in typed handlers.
// When this is supported, the handler will not do internal pooling of objects,
// call Recycle() when the object is no longer needed.
// The recycler should handle nil pointers.
type Recycler interface {
	Recycle()
}

// MSS is a map[string]string that can be serialized.