https://github.com/tsegismont/graphql-server-benchmark

Benchmark of a simple GraphQL server on popular spec implementations
https://github.com/tsegismont/graphql-server-benchmark

benchmark graphql graphql-server

Last synced: about 2 months ago
JSON representation

Benchmark of a simple GraphQL server on popular spec implementations

• Host: GitHub
• URL: https://github.com/tsegismont/graphql-server-benchmark
• Owner: tsegismont
• License: apache-2.0
• Created: 2019-05-22T14:40:31.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2023-12-05T22:22:49.000Z (about 1 year ago)
• Last Synced: 2024-10-20T08:58:20.825Z (2 months ago)
• Topics: benchmark, graphql, graphql-server
• Language: Java
• Homepage:
• Size: 211 KB
• Stars: 14
• Watchers: 6
• Forks: 4
• Open Issues: 3
• Metadata Files:
  • Readme: README.adoc
  • License: LICENSE

Awesome Lists containing this project

README

        
= GraphQL server benchmark

Benchmark of a simple https://graphql.org/[GraphQL] server on popular spec implementations.

== Lastest results

link:results/round1/results.adoc[Round 1]

== The benchmark

The benchmark consists in a simple blogging GraphQL server:

* fetching json (author profiles) from a remote HTTP server

* reading rows (posts and comments) from a Postgres database

image::graphql-server-benchmark.svg[GraphQL Server Banchmark]

[source]

.Schema file

----

type Post {

    author: Author!

    title: String!

    content: String!

    comments: [Comment!]

}

type Comment {

    author: Author!

    post: Post!

    content: String!

}

type Author {

    firstName: String!

    lastName: String!

    bio: String!

    posts: [Post!]

    comments: [Comment!]

}

type Query {

    posts: [Post!]

    author(id: Int!): Author

}

----

The backend HTTP service is setup with 10 authors.

The Postgres database is setup with:

* 2 posts per author

* 2 comments per post:

** 1 comment from another author

** 1 comment (reply) from the post author

Each framework is tested with different link:queries.adoc[queries].

== Guidelines

The server must:

* accept the `BACKEND_HOST` environment variable for configuration

* accept the `POSTGRES_HOST` environment variable for configuration

The server may:

* cache author profiles per-query

* use HTTP and Postgres pipelining

The server must not:

* cache all posts and comments

* cache all author profiles

== Test environment

* 3 quad-core servers Intel Xeon e5-2637v3 128G RAM (injector, server, backend)

* 1 quad-core server Intel Xeon e5-2637v3 256G RAM (postgres)

* 10 GbE switch

== Contributing

You can add an implementation in the link:implementations/[implementations] directory.

Make sure the name describes the software stack.

For example, _Java/vertx-graphql-java_ means the implementation:

* is written in the Java language

* uses Vert.x as the web server

* uses GraphQL-Java engine

Each implementation must provide:

* a `Dockerfile` (build and run)

* a `set-env.sh` file that defines the `FRAMEWORK_NAME` implementation (must be unique across implementations)

For testing purposes, you can build and run with Docker:

* the link:backend/[backend] server

* the link:postgres/[postgres] database

== License

link:LICENSE[Apache 2.0 License]