Transactional event sourcing using Slick (audit log for free!)
Event sourcing is a great alternative to traditional “CRUD”-type architectures. The central concept is a persistent stream of events, which drives all changes to the read model and running any kind of business logic. There’s a lot of technical and business benefits to such an approach, such as being able to re-create the state of the system at any point in time, or keeping a detailed audit log of all actions.
Typically, implementations of event sourcing are presented using a NoSQL data storage, which is great for many use cases (e.g. using Akka Persistence + Cassandra). However, nothing stops us from using a relational database and SQL! In many applications (especially “enterprise”), this brings many benefits, such as powerful and familiar query capabilities and higher guarantees around data consistency.
In this mainly live-coding talk we’ll see one way of implementing transactional event sourcing using the
slick-eventsourcing micro-framework, introducing the core concepts: command handlers, read model updates and event listeners, and how to use them to build an event-sourced application. We’ll see how Slick’s
DBAction and Scala flexibility makes it possible to provide an elegant DSL to build the system from simple functions with minimum dependencies.