6 Years of Distributed GraphQL in Production
summary
Atlassian builds and runs a distributed GraphQL gateway since 2018 using the Nadle execution engine and a central SDL repo that auto-deploys merged pull requests; the gateway enforces schema linting, introspects subgraphs to detect drift, and scales to >1000 contributors. Key runtime feature is hydration via standard directives (translating IDs to objects), ownership keeps each type to one service, and the team shifts from manual reviews and a custom DSL to automated rules and self‑service governance.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 5:04Distributed GraphQL problem
- 10:52Gateway architecture and design
- 14:28Hydration and schema mapping
- 17:20Lessons learned over six years
- 23:04Client-driven schema design challenges
- 25:58Future of distributed GraphQL
- 28:06Q&A session
related talks
27mThe Internal Lens: GraphQL Gateways From a Different Axis
55mGraphQL from the Ground Up
20mScaling GraphQL on AWS: Production Architecture for High-Volume Data Systems
56mDistributed GraphQL APIs
29mCertified to Scale: How N-able Hit 1300% GraphQL Growth in 12 Months
07mshare how avoid these GraphQL pitfalls
23mHow GraphQL Helped Create Scalability and Stability in the Retirement Space.
50mDistributed GraphQL APIs: Breaking Down Monoliths & Unlocking Agility for Dev Teams