ApolloFlow
Problem
Engineering teams using GraphQL had no good way to visualize their schema federation, track field-level usage, or detect breaking changes before deployment. Existing schema registries were either proprietary or required complex infrastructure.
Approach
Built an open-source schema registry with a visual schema graph, field-level analytics, and a CI-friendly diff checker. The core is a GraphQL federation gateway that collects usage metrics and feeds them into a PostgreSQL time-series store. The visualizer renders the schema as an interactive D3 force-directed graph, color-coded by subgraph ownership.
Result
500+ GitHub stars, adopted by 3 teams in production. Detected 47 breaking schema changes across 6 months of CI runs. Published as a schema registry reference architecture used in two conference talks.
Tech Breakdown
The gateway is a thin Node.js layer that delegates to subgraph services. Field usage data is bucketed into 5-minute windows and stored in PostgreSQL with a TimescaleDB extension for efficient time-series queries. The diff engine compares schema snapshots structurally — not just textually — catching subtle breaking changes like nullability shifts and union membership changes.