KUBERNATES

---

Why you need Kubernetes ?(K8s)

• Scaling

• Automation

• Reliability

• Deployment

• Orchestration

• LoadBalancing

• SelfHealing

• Availability

• Management

• Efficiency

---

⚙️ Control Plane

The control plane is the brain of Kubernetes that manages the entire cluster. It receives user requests, decides what should happen, and ensures the cluster stays in the desired state.

👉 Working:
When you run a command like kubectl apply, the control plane:
Accepts the request
Decides where the app should run
Ensures it keeps running correctly

---

Etcd

Etcd is the database of Kubernetes. It stores all cluster information like pods, nodes, configurations, and states.

👉 Working:

Whenever something changes (like creating a pod), data is stored in etcd
It keeps a consistent copy across multiple nodes
Uses the Raft algorithm so all nodes agree on the same data

Think: “Whatever Kubernetes knows = stored in etcd”

---

API Server

The API Server is the front door of Kubernetes. Every request (CLI, UI, or API) goes through it.

Working:

Receives requests like kubectl create
Validates the request (is it correct?)
Stores the desired state in etcd
Sends responses back to the user

Without API Server → you cannot talk to Kubernetes

---

Scheduler

The Scheduler decides which node will run your pod.

👉 Working:

Watches for new pods without assigned nodes
Checks available nodes

Considers:
CPU & Memory
Policies / constraints
Assigns the best node

It doesn’t create pods — just responsible for watching for newly created pods that have no
                           node assigned and selecting the best, most suitable node for 
                           that pod to run on. 

---

Controller Manager

The Controller Manager ensures everything runs as expected.

👉 Working:

Continuously compares:
Desired state (from etcd)
Actual state (running in cluster)

If mismatch happens:

Creates missing pods
Restarts failed ones
Adjusts replicas

Example:
If 3 pods required but only 2 running → it creates 1 more

---

Cloud Controller Manager (Optional)

Used when Kubernetes runs on cloud platforms. It connects Kubernetes with cloud services.

Working:

Creates load balancers
Assigns external IPs
Manages cloud storage
Talks to cloud provider APIs

---

⚙️ Data Plane

The data plane is where your actual applications run. It lives on worker nodes and handles execution, networking, and communication.

👉 Working:

After the control plane decides what should run and where, the data plane:
Runs the containers
Keeps them alive
Handles traffic between them

Think: Control Plane = Brain, Data Plane = Workers

---

Pods

A pod is the smallest unit in Kubernetes that runs your application.

👉 Working:

Contains one or more containers
Runs on a worker node

---

Kubelet

A node agent that runs on every worker node.

Working:

Communicates with the API Server
Receives pod specifications
Ensures the pod is running as expected
Monitors pod health
Reports status back to the API Server

---

Kube-Proxy

A networking component that runs on each node.

Working:

Maintains network rules
Enables communication
Handles traffic:
Load balancing

⚙️ Container Runtime

The component responsible for running containers.

👉 Working:

Exists on each worker node
Runs containers inside pods
Works together with kubelet to ensure containers are running