Have you encountered the problem of distributed transactions? Tell us about the Saga pattern.

Distributed transactions occur when a single business operation involves multiple independent services or databases, and data consistency between them must be ensured. The classic approach with two-phase commit (2PC) is often complex and poorly scalable.

The Saga pattern is an alternative approach to managing distributed transactions through a sequence of local transactions, each accompanied by a compensating operation in case of rollback. Saga breaks down a global transaction into steps that are executed sequentially. If one of the steps fails, compensating transactions are triggered to undo the already completed steps.

Example:

• Order service creates an order.
• Payment service reserves funds.
• Delivery service plans the delivery.

If, for example, delivery planning fails, compensating operations are triggered: canceling the fund reservation and canceling the order.

Saga can be implemented in two ways:

• Orchestration: a central coordinator manages the execution of steps and compensating transactions.
• Choreography: each service publishes events and reacts to them without a central coordinator.

This pattern helps achieve consistency in distributed systems without locks and complex protocols.