README.md

Containerization & Orchestration

This repository now includes container and Kubernetes deployment assets that support:

• Docker-based packaging
• Kubernetes orchestration
• Horizontal auto-scaling (HPA)
• Rolling updates
• Self-healing pods (liveness/readiness probes)
• Worker isolation (dedicated worker deployment + node scheduling hints)

Docker

Build images

docker build -f Dockerfile.backend -t flexiroaster-backend:local .
docker build -f Dockerfile.frontend -t flexiroaster-frontend:local .

Run locally with Docker Compose

docker compose up --build

Services:

• Frontend: http://localhost:8080
• Backend: http://localhost:8000
• Worker: isolated background worker process

Kubernetes

Kubernetes manifests are under deploy/k8s and can be applied with Kustomize:

kubectl apply -k deploy/k8s

What is included

• backend.yaml: API deployment/service with rolling update strategy and probes.
• frontend.yaml: web deployment/service with rolling update strategy and probes.
• worker.yaml: isolated worker deployment with node selector/tolerations.
• autoscaling.yaml: HPAs for backend and worker.
• namespace.yaml: dedicated namespace.

Managed Kubernetes options

These manifests are cloud-agnostic and can be deployed to:

• AWS EKS
• Google GKE
• Azure AKS

Recommended managed-cluster setup

1. Create separate node pools for API/web and workers.
2. Label/taint worker nodes to enforce isolation:
   • Label: workload=worker
   • Taint: dedicated=worker:NoSchedule
3. Install Metrics Server (or provider equivalent) for HPA.
4. Use a cloud load balancer + Ingress controller for public access.
5. Push images to a cloud registry (ECR/GAR/ACR) and update image references.

Notes

• Replace placeholder image names (ghcr.io/your-org/...) before deployment.
• Consider adding PodDisruptionBudgets, NetworkPolicies, and secrets management for production hardening.