🏗️ High-Level Architecture

I structured the solution using Clean Architecture, separating concerns into distinct layers:

• Api Layer: Handles HTTP endpoints, validation, and request routing.
• Application Layer: Encapsulates domain logic, MediatR commands, and strategies (e.g., balance adjustment).
• Infrastructure Layer: Implements persistence, ECB integration, jobs, caching, and rate limiting.

All dependencies flow inward — lower layers have no references to upper layers, ensuring testability and maintainability.

---

⚙️ Core Implementations

1️⃣ ECB Gateway Library

• Created a dedicated library (EcbRateService) that acts as a gateway between the app and ECB API.
• Uses HttpClient abstraction (IHttpClientHelper) for request resiliency.
• Parses ECB XML data into strongly-typed CurrencyRate entities.
• Updates the DB via a PostgreSQL UPSERT (ON CONFLICT DO UPDATE) for high efficiency.
• Results are cached in-memory (IMemoryCache) for read performance.

✅ Benefit: Separation of concerns, reusable across other services, and faster read operations.

---

2️⃣ Periodic Background Job

• Integrated Quartz.NET to run EcbRateUpdateJob every minute.
• Fetches, parses, and stores ECB currency rates.
• Supports retry/backoff (via Polly) and misfire handling to ensure reliability under network or DB transient errors.

✅ Benefit: High reliability and automatic synchronization of currency data.

---

3️⃣ Wallet Service

• Implements wallet operations (create, get, adjust) with concurrency-safe updates using SemaphoreSlim.
• Uses Strategy Pattern to encapsulate balance operations (AddFunds, SubtractFunds, etc.).
• Employs caching and invalidation to optimize repeated lookups.

✅ Benefit: Safe concurrent updates and easily extendable strategies for new business rules.

---

4️⃣ Caching & Performance

• Rates and wallets are cached in-memory.
• Cache invalidation triggers after balance changes.
• ECB rate cache automatically refreshed after each job run.

✅ Benefit: Reduces DB load and ensures sub-millisecond reads for frequent queries.

---

5️⃣ Rate Limiting

• Configured per-client IP Token Bucket Rate Limiter.
• Allows 30 requests per minute per client.

✅ Benefit: Prevents abuse and ensures fair access to APIs.

---

🧩 Patterns Used

Pattern	Purpose	Benefit
Strategy Pattern	Handle different wallet adjustment algorithms dynamically	Extensible, no if/else clutter
Factory Pattern	Create wallet strategies dynamically by name	Promotes loose coupling
Dependency Injection (Autofac)	Manage lifetimes and dependencies	High testability
Caching	Store frequently accessed data	Boosts performance
Rate Limiting	Control request bursts	Prevents service degradation