- When the cluster size is small and the scheduler doesn't get unscheduled Pods frequently, the scheduler doesn't try to reschedule Pods in some cases. ([#119784](https://github.com/kubernetes/kubernetes/pull/119784), [@sanposhiho](https://github.com/sanposhiho)) [SIG Scheduling and Testing]

## Dependencies

### Added
_Nothing has changed._

    if (plan is FailedPlan) return plan.result

    // Connect TCP asynchronously.
    tcpConnectsInFlight += plan
    val taskName = "$okHttpName connect ${routePlanner.address.url.redact()}"
    taskRunner.newQueue().schedule(
      object : Task(taskName) {
        override fun runOnce(): Long {
          val connectResult =
            try {
              plan.connectTcp()
            } catch (e: Throwable) {

            .build()

        val startNanos = System.nanoTime()
        val call = client.newCall(request)

        // Schedule a job to cancel the call in 1 second.
        executor.schedule({
          System.out.printf("%.2f Canceling call.%n", (System.nanoTime() - startNanos) / 1e9f)
          call.cancel()

                      return Thread.currentThread().getStackTrace().length;
                    }
                  },
                  directExecutor()));
        } else {
          // Otherwise, schedule a task on directExecutor
          unused =
              serializer.submit(
                  new Callable<@Nullable Void>() {
                    @Override
                    public @Nullable Void call() {

        this.asyncNotifyTasks = new ConcurrentHashMap<>();
        this.scheduler = Executors.newScheduledThreadPool(3); // Increased for async notifications
        this.rpcClient = rpcClient;

        // Schedule periodic tasks
        schedulePeriodicTasks();
    }

    /**
     * Schedules periodic tasks for heartbeat and registration monitoring.
     */
    private void schedulePeriodicTasks() {

- **Cluster Lifecycle**
  - [alpha] Ensure critical cluster infrastructure pods (Heapster, DNS, etc.) can schedule by evicting regular pods when necessary to make the critical pods schedule. ([docs](http://kubernetes.io/docs/admin/rescheduler/#guaranteed-scheduling-of-critical-add-on-pods)) ([kubernetes/features#62](https://github.com/kubernetes/enhancements/issues/62))

        this.leaseToPath = new ConcurrentHashMap<>();
        this.scheduler = Executors.newScheduledThreadPool(1);
        this.changeNotifier = new DirectoryChangeNotifier(this);

        // Schedule periodic cache cleanup
        scheduler.scheduleAtFixedRate(this::cleanupExpiredEntries, 60, 60, TimeUnit.SECONDS);
    }

    /**
     * Request a directory lease
     *
     * @param directoryPath directory path

        // Resource state changed - may need to reconnect
        log.info("Resource change detected for: {}", notification.getResourceName());

        // Schedule reconnection attempt
        scheduleReconnection(getContext().getConfig().getWitnessReconnectDelay());
    }

    /**
     * Handle client move events
     */

        this.cleanupExecutor = Executors.newSingleThreadScheduledExecutor(r -> {
            Thread t = new Thread(r, "LeaseManager-Cleanup");
            t.setDaemon(true);
            return t;
        });

        // Schedule periodic cleanup
        long cleanupInterval = context.getConfig().getLeaseTimeout();
        if (cleanupInterval <= 0) {
            cleanupInterval = DEFAULT_LEASE_CLEANUP_INTERVAL;
        }

service will be automatically spread across zones, enabling applications to
tolerate zone failure.
     * Simplified way to run a container on every node (DaemonSet API (Beta) in the
Extensions API group): Kubernetes can schedule a service (such as a logging
agent) that runs one, and only one, pod per node.
     * TLS and L7 support (Ingress API (Beta) in the Extensions API group): Kubernetes