mon.Start(stop)

	go func() {
		updateTicker := time.NewTicker(100 * time.Millisecond)
		numUpdates := 10
		for {
			select {
			case <-stop:
				updateTicker.Stop()
				return
			case <-updateTicker.C:
				mon.updateCh <- struct{}{}
				numUpdates--
				if numUpdates == 0 {
					updateTicker.Stop()
					return
				}
			}
		}
	}()

	// Simulate GuaranteedUpdate behavior (retries internally on etcd changes if the incoming resource doesn't pin resourceVersion)
	for {
		currentPod := p.startingPod
		if p.numUpdates > 0 {
			currentPod = p.updatePod
		}
		p.numUpdates++

		// Remember the current resource version
		currentResourceVersion := currentPod.ResourceVersion

		obj, err := objInfo.UpdatedObject(ctx, currentPod)
		if err != nil {

		}
	}
}

func verifyUpdates(t *testing.T, manager *manager, expectedUpdates int) {
	t.Helper()
	// Consume all updates in the channel.
	numUpdates := manager.consumeUpdates()
	if numUpdates != expectedUpdates {
		t.Errorf("unexpected number of updates %d, expected %d", numUpdates, expectedUpdates)
	}
}

func (m *manager) consumeUpdates() int {
	updates := 0
	for {
		select {
		case <-m.podStatusChannel:

          index_depth == 1 && num_updates == 1) {
        ImplicitLocOpBuilder builder(loc, rewriter);
        auto indices_shape = BuildIntArrayConstOp(
            builder, rewriter,
            llvm::SmallVector<int64_t>({num_updates, index_depth}),
            rewriter.getI32Type());
        new_indices = rewriter.create<TF::ReshapeOp>(
            loc,
            RankedTensorType::get({num_updates, index_depth},

	handleErr(t, err, "endpointSliceChangesUpdated")
	if actualUpdated != float64(em.numUpdated) {
		t.Errorf("Expected endpointSliceChangesUpdated to be %d, got %v", em.numUpdated, actualUpdated)
	}

	actualDeleted, err := testutil.GetCounterMetricValue(metrics.EndpointSliceChanges.WithLabelValues("delete"))
	handleErr(t, err, "desiredEndpointSlices")

will be nondeterministic if `indices` contains duplicates.

`indices` is an integer tensor containing indices into a new tensor of shape
`shape`.

* `indices` must have at least 2 axes: `(num_updates, index_depth)`.
* The last axis of `indices` is how deep to index into `tensor` so  this index
  depth must be less than the rank of `tensor`: `indices.shape[-1] <= tensor.ndim`