Aggregate Event Stream Lifetimes

Types of aggregates

In an event-sourced system, there are two kinds of aggregates:

• Short-lived aggregates

• Long-lived aggregates

Short-lived aggregates

Short-lived aggregates will not see many events in their short lifetime. Examples of a short-lived aggregate would be Order from the Ordering module and Basket from the Shopping Baskets module. Both exist for a short amount of time, and we do not expect them to see many events.

The performance of short-lived aggregates and streams with few events in general will not be a problem. The small number of events can be read and processed quickly.

Long-lived aggregates

Long-lived aggregates will see many events over their very long lifetime. Examples of long-lived aggregates are Store from the Store Management module and Customer from the Customers module. These entities will be around for a long time and can end up seeing many events.

Larger streams would take longer to read and process; the larger it is, the longer it would take.

Taking periodic snapshots of the event stream

When we know that we will be dealing with a larger stream, we can use snapshots to improve performance by reducing the number of events we will load and process. In the below figure, the state of the stream is saved along with the aggregate version.