Code Challenge - Sergio Sixi

[sergiosixi]

Notes

To solve the code challenge, two microservices were implemented using NestJS, Postgres, Kafka with Schema Registry, and Jest. The services were designed following Hexagonal Architecture and Domain-Driven Design, and are ready to support new use cases.

anti-fraud-ms

Handles the incoming Kafka messages from the "transactions.created" topic, and perform the business rule mentioned in the challenge. It emit two events: "transactions.rejected" and "transactions.approved" depending on the validation result.

transactions-ms

Expose a GraphQL API to create new transactions validated with class-validator package and to retrieve all of them just to take a look to the db. The transactions are store in a Postgres database. In order to handle massive request of write and read, the CQRS pattern were implemented using NestJS libraries. Despite this, using a MongoDB database for read operations, along with synchronization events, could have been a better approach.

This microservice emit messages to the "transactions.created" topic after saving a new transaction in the database. It also instance a Kafka consumer to handle the "transactions.rejected" and "transactions.approved" events events in order to update the transaction status.

Setup

The entire project is organized in a monorepo and can be run using the docker-compose up -d command.

Testing

In order to test the solution, you can use the graphql apollo local UI. http://localhost:3000/graphql and send this example payload

{
  "createTransactionInput": {
    "accountExternalIdDebit": "0c19d290-b5a1-4c62-9a33-d8e7bc3e4a02",
    "accountExternalIdCredit": "a3b52c13-787e-4936-b2ef-083d10f7f6b4",
    "tranferTypeId": 2,
    "value": 90
  }
}