Pattern summary

What

A design approach that handles failures in a sequence of operations by executing defined undo or rollback actions to revert the system to a stable state.

WAFSupports resilience and consistency in distributed systems where traditional ACID transactions aren't feasible, enabling eventual consistency with recoverability.

Used with

Long-running, multi-step workflows across services, such as order processing, financial operations, or any business process where steps may fail partway.

WAFEnhances operational excellence by isolating failure handling per step, and supports reliability through controlled compensation logic.

When

When operations can't be committed atomically due to distributed service boundaries and need a way to roll back only the affected parts of the process.

WAFPromotes reliability and fault tolerance by minimizing the blast radius of failures and supporting loosely coupled service designs.

Not Suitable For

Scenarios requiring strong consistency and atomicity across multiple steps, or where compensating logic is complex, risky, or impossible to implement.

WAFMay impact reliability and maintainability if compensations are inconsistent, side-effect-prone, or hard to define; operational excellence depends on strong monitoring.