Event Sourcing - You are doing it wrong
summary
Argues that real event sourcing works when events are stored one stream per aggregate (streams usually contain tens not millions of events) and read models are only added when needed; proves correctness with optimistic concurrency control by comparing last-event versions; warns Kafka lacks the required offset compare today and recommends EventStore and simple JSON schema with schema‑on‑read; shows practical operations: snapshotting via copy‑transform per year, splitting streams to remodel, tombstones for GDPR deletes, and encrypt‑and‑throw‑away‑key as a deletion pattern.
AI-written summary · about
chapters (7)
- 0:00Start
- 1:52Typical misconceptions
- 3:01Event sourcing bootcamp
- 8:14Read models
- 14:07Handlers are often trivial
- 33:57Dealing with errors
- 37:09Key administration
related talks
52mEvent Sourcing for the Cloud Developer
48mEvent Sourcing - what could possibly go wrong?
54mThings I wish I knew before I started with event sourcing
43mUnraveling the Threads: Event-Sourcing, CQRS, and Event-Driven Architecture
51mMicroservices Data Patterns: CQRS & Event Sourcing
52mEvent Sourcing – the what, why & how
52mEvent Sourcing: deep dive
40mEvent Sourcing: the good, the bad and the complicated