1. By installing a separate, very basic "CNI plugin" binary onto the node to forward low-level pod lifecycle events (CmdAdd/CmdDel/etc) from whatever node-level CNI subsystem is in use to this node agent for processing via socket.
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.

    taint:
      # Controls whether or not the untaint controller is active
      enabled: false
      # What namespace the untaint controller should watch for istio-cni pods. This is only required when istio-cni is running in a different namespace than istiod
      namespace: ""

    affinity: {}

    tolerations: []

    cpu:
      targetAverageUtilization: 80
    memory: {}
      # targetAverageUtilization: 80

					return fmt.Errorf("failed to check if namespace is labeled ambient: %v", err)
				}
				if !namespaceIsLabeledAmbient {
					fmt.Fprintf(cmd.OutOrStdout(), "Warning: namespace is not enrolled in ambient. Consider running\t"+
						"`"+"kubectl label namespace %s istio.io/dataplane-mode=ambient"+"`\n", ns)
				}
			}
			gw, err := makeGateway(true)
			if err != nil {
				return fmt.Errorf("failed to create gateway: %v", err)

                          items:
                            type: string
                          type: array
                        credentialName:
                          description: For gateways running on Kubernetes, the name
                            of the secret that holds the TLS certs including the CA
                            certificates.
                          type: string

	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())

When fetching certificates, ztunnel will authenticate to the CA with its own identity, but request the identity of another workload.
Critically, the CA must enforce that the ztunnel has permission to request that identity.
Requests for identities not running on the node are rejected.

	openNetns := s.currentPodSnapshot.Take(string(pod.UID))
	if openNetns == nil {
		log.Warn("failed to find pod netns")
		return fmt.Errorf("failed to find pod netns")
	}
	// pod is removed from the mesh, but is still running. remove iptables rules
	log.Debugf("calling DeleteInpodRules.")
	if err := s.netnsRunner(openNetns, func() error { return s.iptablesConfigurator.DeleteInpodRules() }); err != nil {

-include .istiorc.mk

# Environment for tests, the directory containing istio and deps binaries.
# Typically same as GOPATH/bin, so tests work seamlessly with IDEs.

export ISTIO_BIN=$(GOBIN)

# If we are running in the Linux build container on non Linux hosts, we add the
# linux binaries to the build dependencies, BUILD_DEPS, which can be added to other targets
# that would need the Linux binaries (ex. tests).
BUILD_DEPS:=

                          items:
                            type: string
                          type: array
                        credentialName:
                          description: For gateways running on Kubernetes, the name
                            of the secret that holds the TLS certs including the CA
                            certificates.
                          type: string

    enableCRDTemplates: false

    # Validation webhook configuration url
    # For example: https://$remotePilotAddress:15017/validate
    validationURL: ""
    # Validation webhook caBundle value. Useful when running pilot with a well known cert
    validationCABundle: ""

    # For istioctl usage to disable istio config crds in base
    enableIstioConfigCRDs: true

  defaultRevision: "default"
  experimental: