Kubernetes Architecture & Pod Lifecycle Deep Dive

Kubernetes Architecture Overview

A production Kubernetes cluster consists of two distinct planes: the Control Plane (managing cluster state) and Worker Nodes (running actual containerized workloads).

   ┌──────────────────────────────────────────────────────────┐
   │                   CONTROL PLANE (Master)                │
   │                                                          │
   │   ┌─────────────┐   ┌──────────────┐   ┌─────────────┐   │
   │   │  API Server │◄──┤  Controller  │   │  Scheduler  │   │
   │   │  (kube-api) │   │   Manager    │   │             │   │
   │   └──────▲──────┘   └──────────────┘   └─────────────┘   │
   │          │                                               │
   │   ┌──────▼──────┐                                        │
   │   │ etcd (Raft) │ (Consistent Key-Value Store)           │
   │   └─────────────┘                                        │
   └──────────┬───────────────────────────────────────────────┘
              │ TLS Communication
   ┌──────────▼───────────────────────────────────────────────┐
   │                   WORKER NODE                           │
   │                                                          │
   │   ┌─────────────┐   ┌──────────────┐   ┌─────────────┐   │
   │   │   kubelet   │   │  kube-proxy  │   │ ContainerD  │   │
   │   └──────▲──────┘   └──────────────┘   └──────▲──────┘   │
   │          │                                    │          │
   │   ┌──────▼────────────────────────────────────▼──────┐   │
   │   │  Pod (Container 1)    Pod (Container 2)          │   │
   │   └──────────────────────────────────────────────────┘   │
   └──────────────────────────────────────────────────────────┘

---

Control Plane Components Explained

1. kube-apiserver: The front door for all administrative requests and node synchronization. Only component that talks directly to etcd.
2. etcd: Highly-available, distributed key-value store using the Raft consensus algorithm. Stores all cluster state and secrets.
3. kube-scheduler: Watches for newly created Pods with no assigned node and selects the best node based on resource requests, taints, and affinity rules.
4. kube-controller-manager: Runs core control loops (Node Controller, ReplicaSet Controller, EndpointSlice Controller, ServiceAccount Controller).

---

Pod Lifecycle & Probes

Kubernetes relies on three distinct container probes to ensure application health:

1. Startup Probe

Determines if the application inside the container has initialized. All other probes are disabled until the startup probe succeeds.

2. Liveness Probe

Checks if the container is still running properly. If it fails, kubelet terminates and restarts the container according to its restartPolicy.

3. Readiness Probe

Checks if the container is ready to accept user network traffic. If it fails, the Pod's IP is removed from all Service endpoints.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: cloud-api-service
  labels:
    app: cloud-api
spec:
  replicas: 3
  selector:
    matchLabels:
      app: cloud-api
  template:
    metadata:
      labels:
        app: cloud-api
    spec:
      containers:
      - name: api
        image: ghcr.io/org/cloud-api:v2.4.0
        ports:
        - containerPort: 8080
        resources:
          limits:
            cpu: "500m"
            memory: "512Mi"
          requests:
            cpu: "100m"
            memory: "128Mi"
        livenessProbe:
          httpGet:
            path: /healthz
            port: 8080
          initialDelaySeconds: 15
          periodSeconds: 10
        readinessProbe:
          httpGet:
            path: /ready
            port: 8080
          initialDelaySeconds: 5
          periodSeconds: 5

---

Essential Kubectl Commands for Day-2 Operations

# Get resource usage across nodes
kubectl top nodes

# Stream logs with container prefix
kubectl logs -f deployment/cloud-api-service --all-containers=true

# Execute an interactive debugging shell inside a Pod
kubectl exec -it pod-name -c container-name -- /bin/sh

# View events sorted by timestamp
kubectl get events --sort-by='.metadata.creationTimestamp'