From REST to distributed GQL
summary
Shows a practical migration path from REST to GraphQL: a two‑month mobile proof‑of‑concept that cut payloads from 300KB to 50KB and cut server requests by ~30%, followed by a 6–7 month rollout that replaces endpoints with GraphQL until ~80% coverage. Describes operational controls (DataLoader, query depth limits, disabled introspection, published‑queries whitelist via Apollo Engine) and moving from monolith stitching to Apollo Federation with per‑service entities, a schema registry and rollout checks to avoid breaking changes.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:25REST to GQL reasons
- 3:57REST vs GQL benefits
- 6:28Testing GQL on mobile
- 8:37Scaling GQL across project
- 10:28Migration results to GQL
- 12:10GQL microservice problems
- 14:18GQL Mesh attempt
- 18:52Apollo Federation emergence
- 22:03How Apollo Federation works
- 28:30Federation usage errors
- 31:41Service failures and rejection
- 33:49Federation migration lessons
- 38:07Results with Apollo Federation
- 40:34Summary
related talks
22mWhy Migrate from REST to GraphQL?
29mFrom REST to GraphQL
46mMoving to GraphQL
32mHow to get started and build something with GraphQL
30mWhy GraphQL vs. Why not GraphQL
24mGraphQL in Production: Empowering Efficient Data Retrieval and Unleashing Developer Productivity
35mGraphQL in the Enterprise
46mFrom REST to GraphQL: A 20-Minute Adventure