Yape Code Challenge 🚀

Our code challenge will let you marvel us with your Jedi coding skills 😄.

Don't forget that the proper way to submit your work is to fork the repo and create a PR 😉 ... have fun !!

• Problem
• Tech Stack
• Send us your challenge

---

Problem

Every time a financial transaction is created it must be validated by our anti-fraud microservice and then the same service sends a message back to update the transaction status. For now, we have only three transaction statuses:

1. pending
2. approved
3. rejected

Every transaction with a value greater than 1000 should be rejected.

  flowchart LR
    Transaction -- Save Transaction with pending Status --> transactionDatabase[(Database)]
    Transaction --Send transaction Created event--> Anti-Fraud
    Anti-Fraud -- Send transaction Status Approved event--> Transaction
    Anti-Fraud -- Send transaction Status Rejected event--> Transaction
    Transaction -- Update transaction Status event--> transactionDatabase[(Database)]

Loading

---

Tech Stack

1. Java 21 con Spring Boot 4.x + WebFlux (Reactivo)
2. PostgreSQL 17 como base de datos
3. Kafka para comunicación asíncrona entre microservicios
4. Redis como capa de caché para alto volumen
5. R2DBC para conexión reactiva a PostgreSQL

Getting Started

Prerrequisitos

• Docker y Docker Compose instalados
• Java 21 (solo para desarrollo local)
• Maven 3.9+ (solo para desarrollo local)

Levanta todo con un solo comando

docker-compose up --build

Microservicios disponibles:

Servicio	URL / Puerto
Transaction API	http://localhost:8080
PostgreSQL	localhost:5432
Kafka	localhost:9092
Redis	localhost:6379

---

APIs Documentation

Crear una transacción

POST /v1/transactions
Content-Type: application/json

Probar

curl --location 'http://localhost:8080/v1/transactions' \
--header 'Content-Type: application/json' \
--data '{
    "accountExternalIdDebit": "d62a2a4c-65ac-4c61-aacf-59ab55e7c970",
    "accountExternalIdCredit": "d62a2a4c-65ac-4c61-aacf-59ab55e7c971",
    "transferTypeId": 1,
    "value": 100.10
}'

**Response `201 Created`:**
```json
{
    "transactionAmount": 100.10,
    "transactionExternalId": "578fa71b-f0f1-490b-8a1c-0645fadd2ba4",
    "transactionId": "498329fd-ee36-4525-8b71-f09e56b74a8b"
}

---

Consultar una transacción

GET /v1/transactions/{transactionExternalId}

curl --location 'http://localhost:8080/v1/transactions/578fa71b-f0f1-490b-8a1c-0645fadd2ba4'

Response 200 OK:

{
    "transactionExternalId": "578fa71b-f0f1-490b-8a1c-0645fadd2ba4",
    "transactionType": {
        "name": "Debit"
    },
    "transactionStatus": {
        "name": "Approved"
    },
    "value": 100.10,
    "createdAt": "2026-02-21 16:38:46"
}

---

Kafka Events

Topic	Productor	Consumidor
transaction.created	Transaction Service	Anti-Fraud Service
transaction.status.updated	Anti-Fraud Service	Transaction Service

transaction.created

{
  "transactionExternalId": "578fa71b-f0f1-490b-8a1c-0645fadd2ba4",
  "value": 120
}

transaction.status.updated

{
  "transactionExternalId": "578fa71b-f0f1-490b-8a1c-0645fadd2ba4",
  "status": "Approved"
}

---

Optional Solution

Para manejar alto volumen de lecturas y escrituras simultáneas se implementó una estrategia de Cache-Aside con Redis:

¿Cómo funciona?

Cuando crea la transacción, en simultane ejecuta el evento para kafka y guarda los datos de la transaccion en cache, cuando alguien quiere consultar el endoint del detalle de la transaccion, irá primero a cache y ya no a la base de datos, si no existe, irá a la base de datos. Esto en un escensario de alto volumen, beneficia en cuanto tiempos de respuesta. Una vez que el microservicio antifraude, valida la regla para aprobar o rechazar, envia al topico para que el microservicio de transaction actualice el estado y ademas refresca el cache con el estado tambien.

Send us your challenge

When you finish your challenge, after forking a repository, you must open a pull request to our repository. There are no limitations to the implementation, you can follow the programming paradigm, modularization, and style that you feel is the most appropriate solution.

If you have any questions, please let us know.