Você pode definir tarefas em segundo plano para serem executadas *após* retornar uma resposta.

Isso é útil para operações que precisam acontecer após uma request, mas que o cliente não precisa realmente esperar a operação terminar antes de receber a resposta.

Isso inclui, por exemplo:

* Notificações por e-mail enviadas após realizar uma ação:

	var err error
	funcName := e.getFunctionName()
	if aggFnCount == funcName {
		if e.Count.StarArg {
			// Handle COUNT(*)
			e.aggregate.runningCount++
			return nil
		}

		val, err = e.Count.ExprArg.evalNode(r, tableAlias)
		if err != nil {
			return err
		}
	} else {
		// Evaluate the (only) argument
		val, err = e.SFunc.ArgsList[0].evalNode(r, tableAlias)
		if err != nil {
			return err
		}
	}

The duration, in seconds. The value can range from 900 seconds (15 minutes) up to 365 days. If value is higher than this setting, then operation fails. By default, the value is set to 3600 seconds. If no *DurationSeconds* is specified expiry seconds is obtained from *Token*.

| Params        | Value                                              |
| :--           | :--                                                |

	},
	ErrMalformedExpires: {
		Code:           "AuthorizationQueryParametersError",
		Description:    "X-Amz-Expires should be a number",
		HTTPStatusCode: http.StatusBadRequest,
	},
	ErrNegativeExpires: {
		Code:           "AuthorizationQueryParametersError",
		Description:    "X-Amz-Expires must be non-negative",
		HTTPStatusCode: http.StatusBadRequest,
	},
	ErrAuthHeaderEmpty: {

	if err != nil {
		writeErrorResponseJSON(ctx, w, toAdminAPIErr(ctx, err), r.URL)
		return
	}
	// Report the ILMExpiryStats only if at least one site has replication of ILM expiry enabled
	var replicateILMExpiry bool
	for _, site := range info.Sites {
		if site.ReplicateILMExpiry {
			replicateILMExpiry = true
			break
		}
	}
	if !replicateILMExpiry {

                }
            }
        }
    }

    /* Build a connection. Only one thread will ever call this method at
     * any one time. If this method throws an exception or the connect timeout
     * expires an encapsulating TransportException will be thrown from connect
     * and the transport will be in error.
     */

    /**
     * Establishes a connection to the remote endpoint.
     *

Puedes definir tareas en segundo plano para que se ejecuten *después* de devolver un response.

Esto es útil para operaciones que necesitan ocurrir después de un request, pero para las que el cliente realmente no necesita esperar a que la operación termine antes de recibir el response.

Esto incluye, por ejemplo:

* Notificaciones por email enviadas después de realizar una acción:

	// Consequently, if err != nil, reply is always false
	_, err := l.ll.RUnlock(context.Background(), *args)
	return l.makeResp(resp, err)
}

// ForceUnlockHandler - query expired lock status.
func (l *lockRESTServer) ForceUnlockHandler(args *dsync.LockArgs) (*dsync.LockResp, *grid.RemoteErr) {
	resp := lockRPCForceUnlock.NewResponse()

	_, err := l.ll.ForceUnlock(context.Background(), *args)

Clients might rely on the platform certificates and servers might use a private
organization-specific certificate authority.

By default `HeldCertificate` instances expire after 24 hours. Use `duration()` to adjust.

By default server certificates need to identify which hostnames they're trusted for. You may add as

				objectsFailed = float64(mj.KeyRotate.ObjectsFailed)
				bucket = mj.KeyRotate.Bucket
			case madmin.BatchJobExpire:
				objects = float64(mj.Expired.Objects)
				objectsFailed = float64(mj.Expired.ObjectsFailed)
				bucket = mj.Expired.Bucket
			}
			metrics = append(metrics,
				MetricV2{
					Description: MetricDescription{
						Namespace: bucketMetricNamespace,
						Subsystem: "batch",