Day-31 | KUBERNETES ARCHITECTURE USING EXAMPLES | Kubernetes is Easy #devops #k8s #devopscourse

Summary

In this video, Abhishek explains the architecture of Kubernetes by comparing it with Docker, simplifying complex concepts for easy understanding. He breaks down Kubernetes into the control plane (master) and data plane (worker) components, detailing their roles: the API server, scheduler, etcd, controller manager, and cloud controller manager in the master node, and cubelet, kube-proxy, and container runtime in the worker nodes. The video highlights Kubernetes' advanced features such as clustering, auto-scaling, auto-healing, and multi-cloud support. Practical examples and analogies are used to show why each component is necessary, making Kubernetes architecture accessible for learners and interview preparation.

Takeaways

• 😀 Kubernetes is abbreviated as K8s because there are 8 letters between 'K' and 's'.
• 😀 Kubernetes provides four main advantages over Docker: cluster behavior, auto-healing, auto-scaling, and enterprise-level support.
• 😀 The smallest deployable unit in Docker is a container, whereas in Kubernetes it is a Pod, which can wrap one or more containers with advanced capabilities.
• 😀 Each Kubernetes worker node (data plane) has three main components: Kubelet, Kube-proxy, and Container Runtime.
• 😀 Kubelet ensures Pods are running, monitors their health, and communicates with the control plane if action is needed.
• 😀 Kube-proxy handles networking, IP allocation, and basic load balancing for Pods on each worker node.
• 😀 Container Runtime executes the containers within Pods and can be Docker shim, containerd, CRI-O, or other Kubernetes-compatible runtimes.
• 😀 The Kubernetes control plane (master) components include API Server, Scheduler, etcd, Controller Manager, and Cloud Controller Manager.
• 😀 API Server is the heart of Kubernetes, exposing the cluster to users and handling all external requests, while the Scheduler decides on which nodes Pods should run.
• 😀 etcd acts as a key-value store to maintain cluster state, Controller Manager manages controllers like ReplicaSets, and Cloud Controller Manager enables cloud provider integration.
• 😀 Kubernetes architecture allows auto-scaling, auto-healing, and enterprise-level orchestration by separating decision-making (control plane) from execution (data plane).
• 😀 Cloud Controller Manager is optional for on-premise setups and allows adding support for new cloud providers via open-source contributions.
• 😀 Understanding Kubernetes components and their interactions is critical for interviews, practical deployment, and designing resilient containerized applications.

Q & A

Why is Kubernetes sometimes referred to as K8s?

Kubernetes is called K8s because the '8' represents the eight letters between 'K' and 's' in the word 'Kubernetes'. It is a common shorthand used in the DevOps community.

What are the four fundamental advantages Kubernetes offers over Docker?

Kubernetes offers cluster behavior by default, auto-healing of pods, auto-scaling of resources, and multiple enterprise-level features such as advanced load balancing, networking, and security, which Docker alone does not provide.

What is the smallest deployable unit in Docker and Kubernetes, respectively?

In Docker, the smallest deployable unit is a container, whereas in Kubernetes, it is a pod, which can contain one or more containers and provides advanced features like replication and auto-healing.

What are the main components present on a Kubernetes worker node (data plane)?

Each Kubernetes worker node has three main components: Kubelet, which ensures pods are running; Kube-proxy, which handles networking and load balancing; and the Container Runtime, which executes containers inside pods.

What role does the Kubelet play in Kubernetes architecture?

The Kubelet is responsible for managing pods on a worker node. It ensures that pods are running as expected, reports the status to the control plane, and triggers auto-healing actions if a pod fails.

What is the purpose of the API Server in the Kubernetes control plane?

The API Server is the central component of the control plane. It exposes Kubernetes to external users, receives requests, and serves as the entry point for all commands and interactions with the cluster.

How does the Kubernetes Scheduler function within the control plane?

The Scheduler determines which worker node a pod should be deployed on based on resource availability and other constraints. It acts on instructions from the API Server to efficiently distribute workloads.

What is etcd and why is it important in Kubernetes?

etcd is a distributed key-value store that holds the entire state and configuration of the Kubernetes cluster. It is crucial for backup, restoration, and maintaining cluster consistency.

What is the Cloud Controller Manager (CCM) and when is it required?

The Cloud Controller Manager is responsible for translating Kubernetes requests to cloud-specific APIs, such as creating load balancers or storage. It is required only when running Kubernetes on cloud platforms and is optional for on-premise deployments.

What is the role of the Controller Manager in Kubernetes?

The Controller Manager manages built-in controllers, such as the ReplicaSet controller, which ensures that the desired number of pod replicas are running. It monitors cluster state and enforces policies like auto-scaling and state consistency.

How does Kubernetes ensure networking and load balancing for pods?

Kube-proxy on each worker node manages pod networking by assigning IP addresses and handling load balancing. It uses mechanisms like IP tables on Linux to distribute network traffic across pods.

Why does Kubernetes have more components than Docker for running containers?

Kubernetes provides enterprise-level features such as auto-scaling, auto-healing, cluster orchestration, and cloud provider integration. These advanced capabilities require additional components like Kubelet, API Server, Scheduler, Controller Manager, etcd, and CCM, which Docker alone does not need.