GraphQL and Microservices or How we learned to be big and small at the same time
summary
Company moves to a GraphQL gateway over microservices, keeping GraphQL as a thin API layer and pushing business logic and authorization into SDKs and backing services. Uses Apollo client on the frontend, Facebook DataLoader for per-request batching/caching, and standard JSON RPC between services. CDN caching proves ineffective; metrics require Apollo Optics plus request IDs into NewRelic/ELK to find slow services. Builds a React Native read-only app in about two weeks using the existing schema and reports that the approach works and eases migrations.
AI-written summary · about
chapters (14)
- 0:00Introduction and welcome
- 0:25Journey from ad-hoc APIs to GraphQL
- 7:49What GraphQL is and isn't
- 15:13Resolvers and data structure
- 16:20Authentication and authorization
- 18:28Protocol agnosticism and best practices
- 19:31Data loaders and caching
- 24:07Monoliths vs microservices
- 25:29Platform architecture and SDKs
- 28:19SDKs and business logic
- 29:24Benefits of GraphQL at Vivint Solar
- 31:56Mobile app development with GraphQL
- 33:18Metrics and monitoring in GraphQL
- 34:46Client implementations and future
related talks
40mNODEJS & GRAPHQL COOKBOOK. LET’S TALK ABOUT MICRO-SERVICES
51mWhy GraphQL is Perfect for Microservices
23mA glimpse of GraphQL" Frontend developers’ life is easy now!"
46mMoving to GraphQL
55mGraphQL - APIs the New Way
28mGraphQLifying React Native App
27mWhy GraphQL is Perfect For Microservices
01h 13mFrom Monolith to Microservices and Back