E-commerce Microservices – Detailed project overview

Technical overview

Tech stack: React 18 + TypeScript (frontend), Spring Boot 3.5 + Java 21 (backend)
Architecture: Microservices with three independent backend services (User, Product, Order)
Authentication: JWT-based authentication with a centralised user service
Deployment: Docker Compose for local development, Docker Hub for image distribution
Testing: JUnit 5 with JaCoCo coverage reports for each microservice

Architecture & integrations

The project follows microservice principles where each service has its own responsibility and database. Services communicate via REST APIs and can be deployed and scaled independently of one another.

User Service: Handles registration, login, JWT tokens and user permissions
Product Service: Product catalogue with CRUD operations, image handling and stock management
Order Service: Order handling, payments via Stripe and AI-generated horoscopes
Frontend: React SPA with Vite, served via Nginx in Docker
OpenAPI/Swagger: Automatic API documentation for all backend services
Stripe integration: Payment Intents API for secure payments and webhook handling
OpenAI/1min.ai: AI generation of personalised horoscopes based on user data
OpenPDF: Generates PDF documents for digital products (horoscopes)
Spring Actuator: Health checks and metrics for each service

Screenshots

Registration form with fields for personal details and password — 1. Registration where users enter personal details to create an account and are automatically signed in.

Product catalogue with AI horoscopes in different tiers — 2. Product gallery with buy buttons and price information.

Shopping cart with a selected product and buy buttons — 3. The cart showing selected products, quantity and total.

Checkout page with a Stripe payment form — 4. Stripe checkout with card payment and Klarna as an alternative.

Form where customer data generates a personalised horoscope — 5. The AI flow where the customer's details are sent for product generation.

PDF view of a delivered AI horoscope — 6. The delivered horoscope with inline PDF viewing and a download option.

Order history showing a completed order — 7. Order history with status, total and item information.

Key features

User management: Registration, login, profile management and activity logging
Product catalogue: Create, edit and delete products with image upload and categorisation
Stock management: Real-time stock updates with reservation at purchase
Order handling: A complete order flow from cart to confirmation, with order history
Stripe payments: Secure card payments with Payment Intents and webhook validation
Entitlement system: Management of digital rights and consumable products
AI horoscopes: Personalised horoscopes generated with OpenAI and delivered as PDF
Analytics: Order statistics and sales data for administrators
Service-to-service communication: The Order Service calls the User and Product services for validation

Microservice architecture

Each service is fully independent with its own database (H2 for development) and can be deployed separately. This gives the flexibility to scale specific parts of the system based on load.

Separation of concerns: Each service has a clear responsibility without overlap
Independent deployment: Services can be updated without affecting other parts of the system
Scalability: Individual scaling of services based on need (e.g. more Product Service instances)
Resilience: If one service goes down, the whole system is not necessarily affected
Docker Compose: A simple local development environment with all services and networking configured
Version control: Each service has its own Git repository as a submodule

Lessons

This project gave me a deep understanding of microservice architecture and the challenges that come with distributed systems. Designing API contracts between services, handling service-to-service communication and ensuring data consistency across service boundaries requires careful planning. Implementing JWT-based authentication shared across all services taught me the importance of centralised user management and secure token validation.

Working with Stripe webhooks taught me the importance of idempotency and secure webhook validation to avoid double charges. The Docker Compose configuration made it possible to run the whole system locally with realistic network conditions, which sped up development considerably.