The fundamental idea of Event Sourcing is that of ensuring every change to the state of an application is captured in an event object, and that these event objects are themselves stored in the sequence they were applied for the same lifetime as the application state itself.
If you need a basic introduction to event sourcing check Martin Fowlers Event Sourcing description.
Triggering events
- Some method is called on an aggregate causing an event
- The event is dispatched
- The event is applied, the status of the aggregate changes
- The event is put in a stack of non-persistent events
When persisting the aggreate:
- The aggregates events that were stacked are read by the event system implementation
- and stored in the event store
- and flagged as read / applied
- Projections (if there are any) process the events
- Projections generate the read model / fill the database
This describes what happens when an aggregate is read / it's state restored.
- Attempt to read the aggregate
- Check snapshot store
- If snapshot of the aggregate is present, get it from there
- Get remaining events if there are any after the version of the snapshot
- If no snapshot is present read all events from the event store from the beginning
- Replay the events on the aggregate
- Check snapshot store