Bookstore Microservices

This repository implements a bookstore platform as a set of Spring Boot microservices behind a Spring Cloud Gateway, plus a React + Vite frontend. The project supports local development with Docker Compose and production-style deployment on AWS EKS using GitOps (Argo CD).

Architecture at a glance

Services

Service	Port (local)	Responsibility
API Gateway	8080	Single HTTP entry point; routes /auth/**, /api/**, /analytics/** to backend services; CORS
Auth Service	8081	Registration, login, JWT access tokens, refresh-token sessions
User Service	8082	User profile CRUD; internal profile creation via API key
Book Service	8083	Catalog (books, authors, categories, publishers); S3 book cover upload
Order Service	8084	Order queries and cancellation; builds orders from Kafka payment-success events
Payment Service	8087	Stripe PaymentIntent checkout, webhooks, payment state
Notification Service	8085	Email and SMS on order-created (Kafka consumer only — no HTTP API)
Analytics Service	8088	Admin dashboards fed from Kafka events
Frontend	5173 dev / 80 prod	React SPA: catalog, cart, checkout, admin analytics

Repository layout

auth-service/
user-service/
book-service/
order-service/
payment-service/
notification-service/
analytics-service/
api-gateway/
frontend/
docker/
docker-compose.yml
.github/workflows/ci-workflow.yml
bookstore-docs/

Production Kubernetes manifests, Terraform, and Argo CD applications live in the separate bookstore-infra repository.

Key platform capabilities

• User registration and login with JWT + refresh tokens
• JWT-protected customer and admin routes
• Catalog CRUD for books, authors, categories, and publishers
• Book cover upload to Amazon S3 (production)
• Cart and checkout flow in the frontend
• Stripe payment intent creation and webhook handling
• Order creation from payment success events via Kafka
• Notification fan-out (email + SMS) from order events
• Admin analytics views backed by Kafka-fed aggregates
• Docker-based local full-stack environment
• CI/CD: GitHub Actions → Amazon ECR → GitOps manifest update → Argo CD sync

Technology stack

Backend

• Java 17
• Spring Boot 4.x
• Spring Security
• Spring Data JPA
• Spring Cloud OpenFeign
• Spring Cloud Gateway (WebFlux)
• Spring Kafka
• MySQL
• JWT via io.jsonwebtoken
• Stripe Java SDK
• AWS SDK (S3) in book-service
• Micrometer + Spring Actuator (Prometheus metrics on most services)

Frontend

• React 19
• TypeScript 6
• Vite 8
• React Router DOM 7
• Recharts
• @stripe/react-stripe-js

Infrastructure (bookstore-infra)

• Amazon EKS, ECR, RDS (MySQL), S3, IAM, VPC
• Kubernetes manifests in k8s/
• Argo CD with automated sync
• Prometheus, Grafana, Alertmanager in monitoring namespace
• Terraform modules for VPC, EKS, RDS, ECR, S3, GitHub OIDC

Messaging

Local and production environments use Apache Kafka 3.9.1 in KRaft mode (no Zookeeper).

Topic	Producer	Consumers
payment-success	payment-service	order-service
payment-failed	payment-service	analytics-service
order-created	order-service	notification-service, analytics-service

:::warning Known implementation gap payment-service publishes payment-success, while analytics-service also listens for payment-completed (no producer exists for that topic). Successful payment analytics will not be ingested until the topic names are aligned. :::

Deployment modes

Mode	Where configured	Purpose
Local	docker-compose.yml in this repo	Developer full-stack on localhost
Production	bookstore-infra repo (k8s/, Terraform)	AWS EKS cluster with Argo CD GitOps

See Deployment Overview and AWS Architecture Overview for the production pipeline.