Hive

Hive Console registers GraphQL schemas and reports on how they are used, whether the API is federated, stitched or standalone. It closes the gap between publishing a schema and knowing what that change did, because what counts as a breaking change is decided from real operation data rather than a fixed rule.

The CLI publishes schemas and runs checks from any CI system, and the registry is served through a multi-zone Cloudflare CDN. Check results come back through Slack, GitHub or the application itself, the GitHub app can add status checks to pull requests, and the full history of changes stays available even for a composed schema.

Around the registry sit the pieces that show how a deployed API behaves.

• Usage monitoring: error rates and distribution, latency, request volume, operation counts and active clients.
• OpenTelemetry tracing: distributed traces beside usage data, with HTTP metadata and federation subgraph activity.
• App deployments: versioned persisted-document artifacts that can be activated, retired and served through the CDN.
• Alerting: Slack, Microsoft Teams or custom webhooks for schema and metric alerts.
• Governance: organization OIDC and audit log exports for teams that centralize access management.

The stack is Node.js and TypeScript with Fastify and GraphQL Yoga behind the API, React and Vite in the console, and PostgreSQL, Redis, ClickHouse and Kafka underneath. Run it as a managed service, or on your own infrastructure with the community Docker Compose configuration. It is MIT licensed and built in public by The Guild.