sds-flow.svg...

		return c.labelBrokenPod(pod)
	}
	return nil
}

// repairPod actually dynamically repairs a pod. This is done by entering the pods network namespace and setting up rules.
// This differs from the general CNI plugin flow, which triggers before the pod fully starts.
// Additionally, we need to jump through hoops to find the network namespace.
func (c *Controller) repairPod(pod *corev1.Pod) error {
	m := podsRepaired.With(typeLabel.Value(repairType))

This component accomplishes that in the following ways:

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:

    # Kubernetes >=v1.11.0 will create two PriorityClass, including system-cluster-critical and
    # system-node-critical, it is better to configure this in order to make sure your Istio pods
    # will not be killed because of low priority class.
    # Refer to https://kubernetes.io/docs/concepts/configuration/pod-priority-preemption/#priorityclass
    # for more detail.
    priorityClassName: ""

    proxy:
      image: proxyv2

Additionally, using the redirection mechanism reduces the need for clients to know the destination's ztunnel address.

Below shows an example outbound request. The "target" path is what the client sends, while the "actual" path is the real network flow after redirection.

```mermaid
graph LR
    subgraph Client Node
        Client
        CZ["Ztunnel"]
    end
    subgraph Server Node
        Server
        SZ["Ztunnel"]
    end

// FlowDistinguisherMethod specifies the method of a flow distinguisher.
message FlowDistinguisherMethod {
  // `type` is the type of flow distinguisher method
  // The supported types are "ByUser" and "ByNamespace".
  // Required.
  optional string type = 1;
}

// FlowSchema defines the schema of a group of flows. Note that a flow is made up of a set of inbound API requests with

// FlowDistinguisherMethod specifies the method of a flow distinguisher.
message FlowDistinguisherMethod {
  // `type` is the type of flow distinguisher method
  // The supported types are "ByUser" and "ByNamespace".
  // Required.
  optional string type = 1;
}

// FlowSchema defines the schema of a group of flows. Note that a flow is made up of a set of inbound API requests with

    print("##", "*" * 25)

    for k in sorted(common):
        if q0[k] != q1[k]:
            print("## ", k)
            s0 = yaml.safe_dump(q0[k], default_flow_style=False, indent=2)
            s1 = yaml.safe_dump(q1[k], default_flow_style=False, indent=2)

            print(datadiff.diff(s0, s1, fromfile=args.orig, tofile=args.new))

    return changed + len(added) + len(removed)


def main(args):

The following diagram illustrates the flow of unauthenticated traffic with a `PERMISSIVE` policy:

```mermaid
graph TD;
src[src pod]-->|plaintext port|ztunnel{"ztunnel (L4 policy applied here)"}
ztunnel{ztunnel}-->|TLS|wp{waypoint}
wp-->|mTLS|ztunnel

  # https://kubernetes.io/docs/concepts/storage/volumes/.
  volumeMounts: []

  # Configure this to a higher priority class in order to make sure your Istio gateway pods
  # will not be killed because of low priority class.
  # Refer to https://kubernetes.io/docs/concepts/configuration/pod-priority-preemption/#priorityclass
  # for more detail.