Kubernetes Clusters

Scalable container orchestration for production workloads.

Overview

The Kubernetes Cluster service automates container scheduling, scaling, networking, and service discovery. Each cluster includes control plane components and worker nodes. Supports public and private clusters, auto-scaling, persistent storage, load balancing, and monitoring.

---

Prerequisites

• Cloud console or CLI access
• Virtual network/subnet configured
• (Optional) SSH access for debugging worker nodes
• Container images in accessible registry

---

Step 1: Create Cluster

1. Go to Kubernetes > Clusters
2. Click Create Cluster
3. Configure:

Setting	Description
Cluster name	e.g., prod-cluster-01
Region/Zone	Deployment location
Kubernetes version	Select version
Network/Subnet	Pod and service communication

4. Choose cluster type:

   • Public: Accessible via public endpoint
   • Private: Internal network or VPN only

5. Configure node pool:

   • Instance type and size
   • Node count (min/max for auto-scaling)

6. Click Create Cluster

Provisioning takes several minutes.

---

Step 2: Configure Node Pools

• Navigate to Node Pools within your cluster
• Add pools for different workloads (general, GPU, memory-optimized)
• Enable auto-scaling based on resource utilization
• Set taints and labels for pod scheduling control

---

Step 3: Connect to Cluster

1. Download kubeconfig from dashboard
2. Set config as active:

export KUBECONFIG=~/Downloads/kubeconfig.yaml
kubectl get nodes

Worker nodes should show Ready state.

---

Step 4: Deploy Application

kubectl create deployment webapp --image=nginx:latest
kubectl expose deployment webapp --port=80 --type=LoadBalancer

This creates a Deployment managing NGINX pods and a Service exposing them via external load balancer.

---

Step 5: Configure Networking

Component	Description
Service Types	ClusterIP, NodePort, LoadBalancer
Ingress Controller	NGINX or Traefik for HTTP routing
Network Policies	Control pod-to-pod communication
Private Networking	Internal subnets and load balancers

---

Step 6: Add Persistent Storage

1. Go to Storage > Volumes or use PVCs
2. Create PersistentVolumeClaim:

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: app-data
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 20Gi

3. Mount PVC in Pod spec for persistent data

---

Step 7: Monitoring and Logs

• View CPU, memory, pod status in metrics dashboard
• Stream pod logs:

kubectl logs <pod-name>

• Configure autoscaling:
  • HPA: Horizontal Pod Autoscaler for load-based scaling
  • Cluster Autoscaler: Node-level scaling

---

Step 8: Updates and Maintenance

• Upgrade Kubernetes from dashboard (control plane first, then nodes)
• Rotate credentials and API tokens periodically
• Use rolling updates:

kubectl rollout status deployment/webapp

• Backup manifests and secrets regularly

---

Features Summary

Feature	Description
Managed Control Plane	HA Kubernetes API servers and etcd
Node Pools	Worker groups with specific instance types
Auto-Scaling	Dynamic node adjustment
Public/Private	Choose security and exposure level
Load Balancing	Built-in L4/L7 load balancers
Persistent Storage	Block and file storage
Network Policies	Fine-grained traffic control
Monitoring	Resource metrics and alerting
Upgrades	Seamless version upgrades
RBAC	Role-based access control

---

Troubleshooting

Issue	Cause	Solution
Pods stuck in Pending	No nodes or quota exceeded	Scale up node pool
kubectl timeout	API unreachable	Check endpoint and kubeconfig
LoadBalancer not provisioned	No public IP pool	Allocate IPs or contact admin
Volume mount fails	PVC not bound	Verify PVC and storage class
Autoscaling not triggering	Metrics missing	Install metrics server