Documentation
Gateway
Deployment
Serverless / On the Edge
Google Cloud Platform

Deploying Hive Gateway to Google Cloud Platform

Google Cloud Platform (GCP) is a suite of cloud computing services powered by Google. It is easy to use Hive Gateway with GCP.

Prerequisites

You will first need to install the GCP command-line tool: gcloud. You can find instructions here.

If you already have gcloud installed, make sure it is up to date with gcloud components update.

Create a new project and make sure billing is enabled.

⚠️

Running these examples requires you to have billing enabled on your GCP account. It should not cost more than a few cents, but don’t forget to clean up your project after you are done to avoid unexpected charges.

Cloud Functions

Cloud Functions is a serverless execution environment for building and connecting cloud services. With Cloud Functions, you write simple, single-purpose functions that are attached to events, such as an HTTP request.

It is probably the most straight forward way to deploy a Hive Gateway to GCP.

💡

Before you start, make sure you read the Serverless / On the Edge page.

Installation

npm i @google-cloud/functions-framework @graphql-hive/gateway graphql

Don’t forget to add the main field to your package.json. Google Cloud Functions rely on it to know which file to run.

This example uses ESM syntax, so you should set "type": "module" in your package.json.

Usage

Gateway runtime

You can create a runnable index.js script that will be executed by Google Cloud Run Function to serve requests.

See Bundling Problems for more details about how to load the supergraph and transports option.

import { createGatewayRuntime } from '@graphql-hive/gateway'
import http from '@graphql-mesh/transport-http'
import supergraph from './supergraph.js'
 
export const graphql = createGatewayRuntime({
  // All options available in `gateway.config.ts` configuration can also be passed here.
  supergraph,
  transports: {
    http // For example, http transport is required for subgraphs using standard GraphQL over HTTP.
  }
})

You can now deploy your function with gcloud CLI:

$ gcloud functions deploy graphql --runtime nodejs18 --trigger-http --allow-unauthenticated

You can now test your function by using the URL found in the httpsTrigger.url property returned by the previous command or by using the gcloud CLI:

gcloud functions describe graphql

Cloud Run

Cloud Run is the Platform as a Service by Google. It is straightforward to use Hive Gateway with it.

Installation

Create a new Node project and add Hive Gateway to its dependencies.

npm i @graphql-hive/gateway graphql

This example uses ESM syntax, so you should set "type": "module" in your package.json.

Add a start script to your package.json. Cloud Run needs to know how to start your application.

💡
You can use Gateway CLI as usual with gateway.config.ts
{
  "name": "hive-gateway-cloud-run-guide",
  "version": "1.0.0",
  "type": "module",
  "scripts": {
    "start": "hive-gateway supergraph"
  },
  "dependencies": {
    "graphql": "latest",
    "@graphql-hive/gateway": "latest"
  }
}

You can now deploy to Cloud Run. You can use all default values, except the last one, which allows unauthenticated access to your service.

$ gcloud run deploy --source .

If this is your first time using Cloud Run, enabling the service can take up to a few minutes to be fully effective. If you encounter any 403 Forbidden errors, please wait for 2 minutes and try again.

You can now access your API using the URL provided by gcloud. The default GraphQL endpoint is /graphql.

If you need to use TypeScript or any other tool that requires a build phase, such as code generation, add a Dockerfile to the root of your project so that Cloud Run can build a custom image for you.