GraphQL Lambda Server using graphql-server-lambda from Apollo.

graphql-tools and merge-graphql-schemas are used to generate the schema.

serverless-webpack is used to transform ES6 with Babel and build the lambda.

Clone the repository and install the packages.

git clone https://github.com/boazdejong/serverless-graphql-api
cd serverless-graphql-api
npm install

Run the deploy script to create the Lambda Function and API Gateway for GraphQL. This also creates two DynamoDB tables named artists and songs

npm run deploy

Query the GraphQL server using the GraphiQL.app. If you have Homebrew installed on OSX run

brew cask install graphiql

The following mutations are available in this example.

Create an artist providing the first and last name as arguments. The id will be a generated uuid.

mutation {
  createArtist(first_name: "Billy", last_name: "Crash") {
    id
  }
}

Using the generated id from the artist you can create a song with the following mutation. Also provide a title and duration.

mutation {
  createSong(artist: "99a746e0-0734-11e7-b2fd-45ae0a3b9074", title: "Whatever", duration: 120) {
    id
  }
}

mutation {
  updateArtist(id: "99a746e0-0734-11e7-b2fd-45ae0a3b9074", first_name: "John", last_name: "Ruth") {
    id
    first_name
    last_name
  }
}

mutation {
  updateSong(id: "a8a0a060-071b-11e7-bd09-8562f101f7c2", artist: "99a746e0-0734-11e7-b2fd-45ae0a3b9074", duration: 130, title: "A new title") {
    id
  }
}

{
  songs {
    id
    title
    duration
    artist {
      id
      first_name
      last_name
    }
  }
}

This query will return a result similar to this

{
  "data": {
    "songs": [
      {
        "id": "a8a0a060-071b-11e7-bd09-8562f101f7c2",
        "title": "Whatever",
        "duration": 120,
        "artist": {
          "id": "99a746e0-0734-11e7-b2fd-45ae0a3b9074",
          "first_name": "Billy",
          "last_name": "Crash"
        }
      }
    ]
  }
}

This project also includes an example of capturing table activity with DynamoDB Streams. The record lambda function is triggered by two stream events. One for each table.

In serverless.yml:

record:
  handler: lib/handler.record
  events:
    - stream:
        type: dynamodb
        arn:
          Fn::GetAtt:
            - ArtistsDynamoDbTable
            - StreamArn
        batchSize: 1
    - stream:
        type: dynamodb
        arn:
          Fn::GetAtt:
            - SongsDynamoDbTable
            - StreamArn
        batchSize: 1

The stream is enabled when defining the DynamoDB table in the serverless.yml resources.

StreamSpecification:
  StreamViewType: NEW_AND_OLD_IMAGES