To explore and use this application, you'll need to switch to either the "development" or "deployment" branch. Each branch contains instructions within its README on how to set up and run the application. Follow those guidelines to get started!
LearnTrad API is a microservices-based backend system built to support a trading simulation and technical analysis platform. It provides services such as authentication, customer and admin management, market data handling, fetch quiz generation, top-up system, and trade placing simulation and processing — all designed using Spring Boot and deployed using Kubernetes.
A key feature of this platform is the Quiz Fetching system, where customers receive anonymous, preprocessed time-series data in the form of candlestick charts. The data is intentionally obfuscated — its real asset name, time frame, date and price scale are hidden or altered — to ensure users cannot look up actual historical prices while solving the quiz. This encourages customers to hone their technical analysis skills and think independently. Customers can answer the quiz and store their results for future reference.
This system is developed for educational purposes, and serves as a scalable simulation of real-world trading services and challenges.
App Architecture : Lucid.app
ERD : Eraser.io
Utilizing an API Gateway as the single entry point, all requests are filtered before being routed to the appropriate service. The gateway includes a circuit breaker (Resilience4j) to improve fault tolerance. Authentication and authorization are handled by Keycloak, supporting OAuth 2.0 with RS256 and HS256 signature algorithms. Token invalidation (logout) is managed via Redis, ensuring revoked tokens cannot be reused.
Several services use asynchronous communication via Kafka (with Zookeeper and Kafka UI enabled for monitoring), ensuring the system remains responsive under load.
Full observability is provided through the Grafana stack, including:
- Prometheus for metrics
- Tempo and Zipkin for distributed tracing
- Loki for centralized logging
Comprehensive XAUUSD data is available from 2004 to August 2025, with a minimum resolution of 1 minute. It is stored in TimescaleDB for efficient time-series querying. Other business-related data such as customer profiles, top-ups, and trades are stored in PostgreSQL.
The application integrates with the Twelve Data API to provide real-time market prices. Customers have the ability to access this data securely using access tokens, based on their authorization level.
Users can register, create and update their customer profiles, and retrieve their personal data securely. Admins have full access to view all customer information.
Customers are able to:
- Execute market orders and place pending orders (buy stop, buy limit, sell stop, sell limit)
- Modify existing trades and retrieve their own trade data
- All trades are processed and analyzed by the trade-processor-service to determine their status (running, profit, or loss)
Users may top up their balance, receive dummy payment confirmations, and view their own top-up data. Admins have access to all top-up records across the platform.
One of LearnTrad’s core features is the Fetch Quiz, which delivers anonymized historical candlestick data:
- The entity (e.g., XAUUSD) is hidden
- The timeframe is undisclosed
- Price values are rescaled to prevent matching with public historical data
- The date of the data is hidden
- The goal is to encourage customers to hone their technical analysis skills and think independently without "cheating" by referencing actual historical data.
- Customer can answer the quiz and store their results in the database.
This feature allows customers to access historical candlestick data across multiple timeframes: 1min, 5min, 15min, 30min, 1h, 2h, 4h, 6h, 8h, 12h, 1d, 3d, 1w, 1mo