"@local_xla//xla/service:computation_placer",
        "//tensorflow/core",
        "//tensorflow/core/common_runtime/eager:context",
        "//tensorflow/core/common_runtime/eager:tensor_handle",
        "//tensorflow/core/config:flag_defs",
        "//tensorflow/core/config:flags",
        "//tensorflow/core/data/service:dispatcher_client",
        "//tensorflow/core/data/service:grpc_util",

 * if you are writing a application that needs to aggressively release unused resources you may do
 * so.
 *
 * Shutdown the dispatcher's executor service with [shutdown()][ExecutorService.shutdown]. This will
 * also cause future calls to the client to be rejected.
 *
 * ```java
 * client.dispatcher().executorService().shutdown();
 * ```
 *

topic*               (string)    name of the MQTT topic to publish
username             (string)    MQTT username
password             (string)    MQTT password
qos                  (number)    set the quality of service priority, defaults to '0'
keep_alive_interval  (duration)  keep-alive interval for MQTT connections in s,m,h,d
reconnect_interval   (duration)  reconnect interval for MQTT connections in s,m,h,d

* federation service controller: fixing the logic to update DNS records ([#27999](https://github.com/kubernetes/kubernetes/pull/27999), [@quinton-hoole](https://github.com/quinton-hoole))