In Kubernetes (K8S), the main methods for individual components to achieve high availability include redundant deployments and automated failure recovery mechanisms. The following are the main components of Kubernetes to achieve high availability:

1. kube-apiserver:

- High-availability strategy: Provide a unified entry point to the outside world by deploying multiple API server instances on multiple nodes and using a load balancer (e.g., load balancing provided by a cloud provider or a self-built load balancing solution such as HAProxy, Nginx, etc.).

- Data persistence: all state changes are written to the etcd cluster, and etcd itself needs to be deployed for high availability to ensure data consistency.

2. etcd:

- High Availability Strategy: By running an etcd cluster with an odd number of nodes (usually 3, 5, or 7 nodes), the Raft consistency algorithm ensures that services can continue to be provided even if some of the nodes are down.

- Failure Recovery: When an etcd node fails, the remaining nodes can continue the election and ensure data consistency and availability.

3. kube-scheduler and kube-controller-manager:

- High-availability strategy: each component can run multiple instances on multiple nodes independently, and set up a “leader election” mechanism, i.e., through the shared storage (e.g., etcd-based locking service) to determine which instance is the currently active leader, and the other instances are candidates for switching. The other instances are candidates for switching.

- Automatic failure recovery: when the current leader node is unavailable, the instances on the remaining nodes will compete and obtain the leadership right to take over the work to ensure the continuity of scheduling and service control functions.

4. kubelet and kube-proxy:

- Running on each Node node, they are not the focus of the cross-node high availability design, but can be used to achieve high availability at the application level by monitoring and managing the health of the Node as well as container operation, combined with the definition of resources such as Pod replica set (ReplicaSet), Deployment, and so on.

5. Network components (e.g. Calico, Flannel, etc.):

- Usually there are corresponding high availability solutions at the network level as well, such as multi-node deployment and configuration of redundant routing rules.

6. cloud-controller-manager (if applicable):

- Similar to scheduler and controller-manager, it is also possible to deploy multiple instances on multiple nodes and use the leader election mechanism to ensure that only one instance is active in order to interact with the cloud platform and handle cloud service-related tasks.

In summary, with these strategies, Kubernetes is able to ensure that its key components can quickly recover and keep the cluster as a whole running stably in the event of a failure. At the same time, the cluster's self-healing ability is also reflected in the ability to automatically replace failed pods or adjust resource allocation, making the entire system highly reliable.