Event Sourcing a Node.js Application
summary
A concrete postmortem of adding event sourcing to a Node.js conference registration system: uses Node.js and MongoDB, hydrates read/write models per request at first and fails — the registration run collapses to zero signups — then fixes by caching models, reordering to update models before persisting events, and accepting a non‑scalable single‑process critical section. Shows actual failure modes, the surprising scale of conflict events produced, and links to code and slides on github so the implementation can be inspected.
AI-written summary · about
chapters (13)
- 0:00Start
- 0:51Overview
- 4:00The Classical Approach Relational Database
- 4:21Idea behind Event Sourcing
- 9:12Standard Event Sourcing Scenario
- 10:47The Ticket Quota
- 17:53Working Hypothesis
- 19:49What was going on?
- 20:18The End
- 21:25Caching the Global Event Store
- 22:32Caching the Read Models
- 23:00Problem 11
- 24:27Eliminating Parallelization. Part II
related talks
24mHow To Combine Event Sourcing Architecture In Your Node.js Server
40mEvent Sourcing: the good, the bad and the complicated
48mNodeJS Microservices + CQRS + Event Sourcing — Why Do I Do This?
44mEvent-driven Revolutions: From Monolith to Event Sourcing in Practice
48mEvent Sourcing - what could possibly go wrong?
44mEvent Sourcing - You are doing it wrong
44mEvent Sourcing: Core Principles and Implementation in NestJS
52mEvent Sourcing: deep dive