Forex Service API

A Spring Boot based microservice for real-time currency conversion and exchange rate retrieval.
This project is designed with extensibility in mind and showcases common patterns like retry mechanisms, message queue integration, and clean exception handling.

Features

• Currency conversion (single request)
• Bulk currency conversion (via CSV upload)
• Real-time exchange rate retrieval
• Retry mechanism using RabbitMQ for failed conversions
• Scheduled processing of retry queue
• Global exception handling
• API documentation available via Swagger UI
• Dockerized setup for local development

How to Run

Prerequisites

• Java 21
• Maven
• Docker

Build and Run with Docker Compose

First, build the project locally:

mvn clean package

Then run the application with Docker Compose:

docker compose up --build

This will start:

• The application (accessible on http://localhost:8080)
• RabbitMQ (management console at http://localhost:15672, default username/password: guest/guest)

API Documentation

Once the app is running, Swagger UI is available at:

http://localhost:8080/swagger-ui/index.html

Notes and Design Decisions

Why use Scheduled Retry?

This project uses a retry mechanism with in-memory queue + scheduled processor for failed currency conversions.

Reason:
I am using the free plan of CurrencyLayer, which has strict rate limits and occasionally causes API exceptions due to limited access.
To handle this, failed conversions are sent to a RabbitMQ queue and then processed using a scheduled processor.

This is a simple and effective approach for this project scope.
In a production-grade system (with enterprise subscription or unlimited API access), this could be further improved with:

• Backoff strategies
• Circuit breakers and rate limiters

Why are API keys in application.yml?

For simplicity, in this project I have directly added API keys and settings to application.yml.
In a real-world scenario, sensitive information such as API keys should be handled securely:

• Environment variables
• Secrets management tools (Vault, AWS Secrets Manager, etc.)
• CI/CD pipelines with secure injection

This is an area for improvement and can easily be refactored.

Current Limitations / Improvement Ideas

• Use a persistent queue (RabbitMQ + DLQ + persistence)
• Add more integration tests
• Implement user authentication (OAuth2 / JWT)

Technologies Used

• Java 21
• Spring Boot
• Spring Web
• Spring Scheduling
• Spring Cache
• Spring AMQP (RabbitMQ)
• Feign Client
• Lombok
• Swagger / OpenAPI
• Docker
• Docker Compose

The architecture is intentionally kept clean and extensible, and it is open for further development.