* }
 * }
 *
 * <p>It is important to note that the number of permits requested <i>never</i> affects the
 * throttling of the request itself (an invocation to {@code acquire(1)} and an invocation to {@code
 * acquire(1000)} will result in exactly the same throttling, if any), but it affects the throttling
 * of the <i>next</i> request. I.e., if an expensive task arrives at an idle RateLimiter, it will be

    assertEvents(
        "R0.00, R1.38, R1.13, R0.88, R0.63, R0.50, R0.50, R0.50", // #1
        "U4.50", // #2
        "R0.00, R1.38, R1.13", // #3, after that the rate changes
        "R0.88", // #4, this is what the throttling would be with the old rate
        "R0.34, R0.28, R0.25, R0.25", // #5
        "U4.25", // #6
        "R0.00, R0.72, R0.66, R0.59, R0.53, R0.47, R0.41", // #7

    assertEvents(
        "R0.00, R1.38, R1.13, R0.88, R0.63, R0.50, R0.50, R0.50", // #1
        "U4.50", // #2
        "R0.00, R1.38, R1.13", // #3, after that the rate changes
        "R0.88", // #4, this is what the throttling would be with the old rate
        "R0.34, R0.28, R0.25, R0.25", // #5
        "U4.25", // #6
        "R0.00, R0.72, R0.66, R0.59, R0.53, R0.47, R0.41", // #7

        return -1;
    }

    /**
     * Retrieves a list of access results from the data service for processing.
     * Filters out results that are too recent based on commit margin time and manages crawler throttling.
     *
     * @param cb the consumer to customize the search request
     * @param cleanupTime the time taken for the last cleanup operation
     * @return the list of access results ready for processing
     */

      sink.writeUtf8(name)
      sink.writeUtf8(": ")
      sink.writeUtf8(value)
      sink.writeUtf8("\r\n")
    }
    sink.writeUtf8("\r\n")
    sink.flush()
  }

  /** Returns a sink that applies throttling and disconnecting. */
  private fun Sink.withThrottlingAndSocketEffect(
    policy: MockResponse,
    socketEffect: SocketEffect?,
    expectedByteCount: Long,
    socket: MockWebServerSocket,

- Fixed a 1.35.0 regression by disabling the `SchedulerAsyncAPICalls` feature gate. The regression was in scheduler performance, triggered by API client throttling. ([#135904](https://github.com/kubernetes/kubernetes/pull/135904), [@macsko](https://github.com/macsko)) [SIG Scheduling]

- Disabled `SchedulerAsyncAPICalls` feature gate due to performance issues caused by API client throttling. ([#135903](https://github.com/kubernetes/kubernetes/pull/135903), [@macsko](https://github.com/macsko)) [SIG Scheduling]