Data Consistency in Microservices — Saga Pattern

The article explains the distributed transaction problem in microservices, where a single ACID transaction across services is usually not possible.

The Saga pattern runs multi-step workflows through local transactions in services like Order, Inventory, Payment, and Shipping, then uses compensating actions on failure.

It compares event-driven choreography with centralized orchestration, where a Saga Orchestrator coordinates the overall process.

It also stresses that saga state must be persisted so systems can recover and stay consistent during partial failures.