Safely Evolve Your (Federated) GraphQL Schema with GraphQL Hive
summary
Shows an open-source schema registry (GraphQL Hive) that plugs into CI/GitHub Actions via a Hive CLI to run schema check and schema publish commands, store linear version history, and let humans approve or override failed checks. Demo publishes five federated subgraphs via Docker Compose, composition succeeds, and an Apollo Router fetches the supergraph and serves queries successfully. Usage reporting (SDK or router plugin) records field hits (example: a field shows 12 accesses) and enables conditional breaking changes based on percent-of-traffic over 30 days. Storage uses Postgres, ClickHouse, Redis and S3-compatible CDN.
AI-written summary · about
chapters (11)
- 0:00Introduction to Workshop
- 1:14Challenges in Schema Evolution
- 4:43Breaking Change Examples
- 7:05Automating Schema Checks
- 11:48Introduction to Hive Registry
- 15:26Hive Workflow Demo
- 23:40Integrating Hive with CI/CD
- 33:12Federated Schema Management
- 47:03Usage Reporting and Analytics
- 58:08Conditional Breaking Changes
- 1:02:03Audience Q&A
related talks
20mThe State of GraphQL Federation
32mHow to Not Break Your GraphQL Clients
38mGraphQL Federation in the Wild
21mSafely Merging Subgraphs in a Distributed World
26mGraphQL Anywhere - Our Journey With GraphQL Mesh and Schema Stitching
27mStranger Graphs - Building Federated Workflows with Shared Components
19mSupport unlimited customer experiences with GraphQL
22mGraphQL for Everyone