1. By running as a node-level daemonset that:

- listens for these UDS events from the CNI plugin (which fire when new pods are spawned in an ambient-enabled namespace), and adds those pods to the ambient mesh.
- watches k8s resource for existing pods, so that pods that have already been started can be moved in or out of the ambient mesh.

		sortBy := func(pods []*corev1.Pod) sort.Interface { return podutils.ByLogging(pods) }
		sort.Sort(sortBy(pods))
		return config.NamespacedName(pods[0]), nil
	}
	return types.NamespacedName{}, fmt.Errorf("no pods found")
}

type commonFlags struct {
	// output format (json, yaml or short)
	outputFormat string

	proxyAdminPort int

	configDumpFile string

	node string
}

	registerBooleanParameter(constants.RepairLabelPods, false, "Controller will label pods when detecting pod broken by race condition")
	registerStringParameter(constants.RepairLabelKey, "cni.istio.io/uninitialized",
		"The key portion of the label which will be set by the race repair if label pods is true")

    # see
    # https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/#dns-config
    # This does not apply to gateway pods as they typically need a different
    # set of DNS settings than the normal application pods (e.g., in
    # multicluster scenarios).
    # NOTE: If using templates, follow the pattern in the commented example below.
    #podDNSSearchNamespaces:
    #- global

    # see
    # https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/#dns-config
    # This does not apply to gateway pods as they typically need a different
    # set of DNS settings than the normal application pods (e.g., in
    # multicluster scenarios).
    # NOTE: If using templates, follow the pattern in the commented example below.
    #podDNSSearchNamespaces:
    #- global

	// first add all the pods as empty:
	for uid := range ambientPodUIDs {
		s.currentPodSnapshot.Ensure(string(uid))
	}

	// populate full pod snapshot from cgroups
	return s.scanProcForPodsAndCache(ambientPodUIDs)
}

func (s *NetServer) scanProcForPodsAndCache(pods map[types.UID]*corev1.Pod) error {

	// label the pod for exclusion
	labelsPatch := []byte(fmt.Sprintf(`{"metadata":{"labels":{"%s":"%s"}}}`,
		constants.DataplaneModeLabel, constants.DataplaneModeNone))
	_, err = client.Kube().CoreV1().Pods(pod.Namespace).Patch(ctx, pod.Name,
		types.MergePatchType, labelsPatch, metav1.PatchOptions{})
	assert.NoError(t, err)

	// wait for an update events

	for _, excludeNs := range conf.Kubernetes.ExcludeNamespaces {
		if podNamespace == excludeNs {
			log.Infof("pod namespace excluded")
			return nil
		}
	}

	// Begin ambient plugin logic
	// For ambient pods, this is all the logic we need to run
	if conf.AmbientEnabled {
		log.Debugf("istio-cni ambient cmdAdd podName: %s - checking if ambient enabled", podName)
		podIsAmbient, err := isAmbientPod(kClient, podName, podNamespace)

  egressEntities:
   - kube-apiserver

## PodDisruptionBudget settings
## ref: https://kubernetes.io/docs/concepts/workloads/pods/disruptions/
##
podDisruptionBudget:
  enabled: false
  maxUnavailable: 1

## Specify the service account to use for the MinIO pods. If 'create' is set to 'false'
## and 'name' is left unspecified, the account 'default' will be used.
serviceAccount:
  create: true

	if err != nil {
		return fmt.Errorf("failed to get profile, namespace or enabled components: %v", err)
	}

	// Ignore the err because we don't want to show
	// "no running Istio pods in istio-system" for the first time
	_ = detectIstioVersionDiff(p, tag, ns, kubeClient, iop)
	exists := revtag.PreviousInstallExists(context.Background(), kubeClient.Kube())