Architecture of scalable and resilient NodeJS apps with GraphQL & event-driven serverless
summary
Claim: combine Hasura GraphQL Engine subscriptions and live‑queries with event‑triggered serverless functions to make a reactive, async backend. Built a demo food‑ordering stack: React + Apollo frontend, DB row‑events invoke Google Cloud Functions to validate/pay/assign; subscriptions push status to clients. Demo sustains and recovers from bursts — one run updates 18,200 orders (speaker reports ~40,000 previously) — so it works at scale but requires event plumbing, a GraphQL server with subscriptions, and careful handling of race conditions and failure semantics.
AI-written summary · about
chapters (9)
- 0:00Start
- 1:27Traditional architecture
- 4:49Async architecture
- 10:11Example: Food ordering app
- 11:30Architecture Recap
- 12:47High level overview
- 13:53REST vs GraphQL
- 22:43GraphQL & Serverless architecture
- 30:43Food ordering app - scalable
related talks
21mLive-coding demo
39mBuilding serverless applications with GraphQL
46mBuilding modern web-applications with GraphQL & serverless Ruby
50mBuilding Realtime Serverless APIs with GraphQL
19mBuilding Planet Scale Cloud Native Apps with GraphQL & Distributed SQL
36mDeveloping Serverless Applications with GraphQL
29mBuilding modern web-applications with Hasura's GraphQL engine
39mScaling event-driven systems with Node and GraphQL subscriptions