Is GraphQL a trap?
summary
Argues that GraphQL’s wins become traps when misapplied: teams lose velocity to free-lunch migrations that add latency and ops cost; treating the database as an API exposes unwanted data and hides business logic; public endpoints invite abuse unless guarded by query cost analysis or whitelists; splitting schemas by team creates type collisions and maintenance overhead; versioning whole schemas multiplies work — instead add fields and deprecate old ones and keep a single schema powered by a separate business-logic layer.
AI-written summary · about
chapters (8)
- 0:00Intro
- 0:30The Trap: Free-Lunch Migrations
- 1:16The Trap: A Generic Graph Database
- 1:56The Trap: Queries for Everyone
- 2:40The Trap: Everybody Gets a Schema
- 3:41The Trap: Versions
- 4:35One Schema, One Graph
- 5:42The Ever-Evolving Schema
related talks
23mAn Introduction to GraphQL
51mGraphQL: advantages, challenges, and how to overcome them
01h 11mGraphQL: 'Dangerous' Advice. Why Use GraphQL When There’s REST?
22m8 Years of GraphQL: Unraveling the Trade-Offs
18mIs GraphQL the end of RESTstyle APIs
25mGraphQL, the Domain Way
29mDamn GraphQL - Defending and Attacking APIs
23mBlueprints of Successful GraphQL Architectures