What is Hydra?
Hydra is a lightweight framework written in TypeScript designed to solve scaling challenges for GraphQL endpoints such as automated caching, invalidation of relevant cached queries, built-in security (authorization, rate limiting etc...).
Core aspects
- Self-hosted framework
- Operates natively with Authdog backend (OpenID, Authz)
- Cost effectiveness, 100% Edge Computing based
- Automated mutation based invalidation
- Automated cache segmentation
- Cloudflare opinionated (KV, Durable Objects)
Most Hydra operations can be achieved through Hydra CLI.
How does it work?
Hydra acts as an Edge Gateway for GraphQL, meaning Hydra will generate its own unified GraphQL schema from existing GraphQL schema. Requests sent to a Hydra endpoint will automatically cache the queries payload. Queries have dedicated types with their ids stored as metadata in KV. If a given Mutation returns a type with ids in KV, it'll automatically invalidate any matching query in KV with a similar type.
Setup Process
Assuming hydra CLI is installed in your environment, the initial step is generally to generate a dump of the remote endpoints from a given configuration hydra.config.ts in the current project.
Once this dump has been successfully generated, it'll be accessible in .hydra/schemaRaw.js.
Then this generated file will be consumed by Hydra handlers to dynamically generate a JIT unified GraphQL schema.