Can you explain how to implement a distributed transaction within a microservices architecture?

A distributed transaction in a microservices architecture is a way to ensure data consistency when performing operations that involve multiple services. Since each microservice typically manages its own database, traditional ACID transactions do not work directly.

Main approaches:

1. Two-Phase Commit (2PC) — the coordinating service first asks all participants to prepare for commit, then confirms or rolls back the changes. However, 2PC is complex and reduces performance.

2. Sagas — split the transaction into a sequence of local transactions with compensating actions for rollback. If one of the transactions fails, compensating actions are invoked to undo previous steps.

Example of a saga implementation:

• Service A reserves a product.
• Service B deducts money.
• If money deduction fails, Service A cancels the reservation.

Thus, Sagas provide eventual consistency and are more suitable for microservices.

In Java, frameworks supporting Sagas, such as Axon Framework or Spring State Machine, can be used, or the logic can be implemented manually through events and message queues.