Scaling Schema Cardinality Constructing Types at Runtime
summary
Claims a runtime system that constructs GraphQL types per request to represent millions of per-tenant schemas, using GraphQL Java and a per-request cached type Constructor plus many category-owning type Informer actors. A metaschema encodes static vs dynamic parts; object/enum/union/input types are lazily instantiated, answers introspection, honors the spec, and reduces round trips with a type cache. System reports scaling to millions of schemas and remains spec-compliant.
AI-written summary · about
chapters (12)
- 0:00Introduction to schema scaling
- 1:46Dynamic schema challenges
- 3:08Static vs dynamic types
- 4:54Static types with dynamic parts
- 7:00Dynamic types overview
- 8:24Type categories concept
- 10:12Extended SDL examples
- 13:24Architecture requirements
- 14:27Type Constructor and informers
- 23:35End-to-end workflow
- 26:46Recommendations for adoption
- 28:35Q&A session
related talks
26mDynamically Serving a GraphQL API with Custom Types at Runtime
21mBuilt to Evolve: 13 Years of GraphQL
33mEvolving GraphQL Schemas
47mThe State of GraphQL Federation
49mGraphQL Schema Design — Building Schemas Correctly (version 2)
41mOptimizing Shared Types: How Intuit Streamlined Schema Evolution in a Federated Graph
01h 07mGraphQL: Smooth Schemas and Rough Edges
16mBuild Your GraphQL APIs Faster with Nexus Schema