Building Microservices, 2nd Edition

Chapter 6. Workflow

In the previous two chapters, we’ve looked at aspects of microservices related to how one microservice talks to another. But what happens when we want multiple microservices to collaborate, perhaps to implement a business process? How we model and implement these sorts of workflows in distributed systems can be a tricky thing to get right.

In this chapter, we’ll look at the pitfalls associated with using distributed transactions to solve this problem, and we’ll also look at sagas—a concept that can help us model our microservice workflows in a much more satisfactory manner.

Database Transactions

Generically speaking, when we think of a transaction in the context of computing, we think of one or more actions that are going to occur that we want to treat as a single unit. When making multiple changes as part of the same overall operation, we want to confirm whether all the changes have been made. We also want a way to clean up after ourselves if an error occurs while these changes are happening. Typically, this results in us using something like a database transaction.

With a database, we use a transaction to ensure that one or more state changes have been made successfully. This could include data being removed, inserted, or changed. In a relational database, this could involve multiple tables being updated within a single transaction.

ACID Transactions

Typically when we talk about database transactions, we are talking about ACID transactions. ACID is an acronym ...