Last Updated: June 8, 2026
When one business operation spans multiple services, a saga keeps the system consistent without one global transaction.
A saga breaks the operation into local transactions. Each service commits its own step. If a later step fails, earlier steps are undone through compensating actions.
Sagas trade atomicity for eventual consistency and availability. The cost is writing reliable recovery logic yourself.
This chapter covers saga basics, choreography, orchestration, compensations, non-reversible steps, lack of isolation, and crash recovery.