How did you split a monolith into microservices? What logic did you use to divide by domains? How did you ensure domain independence and manage transactions?

Breaking down a monolith into microservices usually begins with analyzing business domains and identifying contexts (bounded contexts) according to Domain-Driven Design (DDD) principles. The logic of the breakdown:

• Identifying independent business functions or areas of responsibility.
• Minimizing dependencies between services to ensure loose coupling and high cohesion.

To ensure domain independence:

• Each microservice has its own database or schema to avoid direct access to other services' data.
• Interaction occurs through well-defined APIs or messages (event-driven architecture).

Managing transactions in a distributed system:

• Using the Saga pattern to coordinate local transactions with compensating operations.
• Avoiding distributed transactions with two-phase commit due to complexity and performance reduction.

Example: when ordering a product, one service handles the order, another handles payment, and a third manages the warehouse. If the payment fails, the Saga initiates the cancellation of the order and the reservation of the product.