fasibio/graphqldockerproxy repository overview

⁠GraphqlDockerProxy

A generic Graphql Docker Proxy

It's a GraphQL API Gateway.

⁠Features

• Continuously integrates the Backend GraphQL Nodes (No restart!)
• !!!It works with Docker and Kubernetes!!!
• Supports load balancing (with docker)

⁠Run with Docker (Quickstart)

⁠How Does it Work:

It works without dependencies. You can start it in your docker cloud. Use it to manage your GraphQL-Microservices. With docker labels you can registry your microservices in the proxy. The proxy automatically will find your services and add them to the gateway.

⁠How to Start the Proxy

In this example we will use docker-compose to start the proxy. Here is an example docker-compose file:

version: '3'
services: 
  api: 
    restart: always
    image: fasibio/graphqldockerproxy
    expose:
      - 3000
    ports:
      - 3000:3000
    networks:
     - web
    environment: 
      - qglProxyRuntime=dockerWatch
      - dockerNetwork=web
      - gqlProxyToken=1234
    volumes: 
     - /var/run/docker.sock:/var/run/docker.sock
networks:
  web:
    external: true

This will start the proxy on port 3000. It's important to include the docker.sock as volume. You can do this with the following environment variables:

• qglProxyRuntime: can be docker or kubernetes. For Docker you need docker ^^
• dockerNetwork: the network where the backend GraphQL-Server will run
• gqlProxyToken: a token which verifies that the microservice belongs to the proxy

That's all!! Now you can open the proxy playground under http://127.0.0.1:3000/graphql⁠ . The API is reachable under http://127.0.0.1:3000/graphql⁠ too.

At the moment it is an empty gateway.

⁠Let's Start a GraphQL Microservice

It is important to put your microservice in the same network as the proxy (In this example the network is called 'web'). Now you have to set the following labels:

• gqlProxy.token: The same token you set in the proxy. (In this example 1234)
• gqlProxy.url: This is the relative path to the proxy running inside the container. (For example: :9000/graphql)
• gqlProxy.namespace: The namespace that wraps your microservice.

⁠Let's Write an Example Microservice

For this example we will use the Docker image bfillmer/graphql-swapi

Create a docker-compose file:

version: '3'
services: 
  swapi: 
    image: bfillmer/graphql-swapi
    expose:
      - 9000
    networks: 
      - web
    labels: 
      - gqlProxy.token=1234
      - gqlProxy.url=:9000/graphql
      - gqlProxy.namespace=swapi
networks:
  web:
    external: true

Start the docker-compose file. The proxy will automatically find the microservice and include it. Under http://127.0.0.1:3000/graphql⁠ you can now see that swapi has wrapped your graphql microservice Inside this namespace you can make graphql requests. For example:

{
  swapi{
    allFilms{
      films{
        title
      }
    }
  }
}

Or you can use the admin Page to see what has been included http://127.0.0.1:3000/admin/graphql⁠ (See the second tab at playground)

⁠Now Let's Scale the GraphQL Microservice !

The proxy knows how to reference the same images with a round robin loadbalancer.

Go in the folder where the SWAPI service is.

Enter the command: sudo docker-compose scale swapi=3

The proxy will automatically start a loadbalancer

⁠Note

You can find examples (and docker-compose files) in the example directory of this git project.

⁠Run with Kubernetes

Complete Doku will coming soon.

The service is available for Kubernetes. It will use the Kubernetes API to find available GraphQL Endpoints.

General use is the same like docker. See "How it works with Docker". You have to set labels in the Deployment-Manifest.

The following labels a available/necessary

• qglProxyRuntime: can be docker or kubernetes or dockerWatch or kubernetesWatch. For Kubernetes you need kubernetes* ^^
• gqlProxyToken: a token which verifies that the microservice belongs to the proxy
• kubernetesConfigurationKind: How the proxy find the Kubernetes API.
  • fromKubeconfig: A Config file which is mount in the Container
  • getInCluster: The POD as it self.

Here is a Example How to run:

⁠The Yaml for the GraphQL Proxy:

Deployment.yaml

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  annotations:
    kompose.cmd: kompose convert
    kompose.version: 1.13.0 (84fa826)
  creationTimestamp: null
  labels:
    io.kompose.service: api
  name: api
  namespace: gqlproxy
spec:
  replicas: 1
  strategy: {}
  template:
    metadata:
      creationTimestamp: null
      labels:
        io.kompose.service: api
    spec:
      containers:
      - env:
        - name: gqlProxyToken
          value: "1234"
        - name: kubernetesConfigurationKind
          value: getInCluster
        - name: qglProxyRuntime
          value: kubernetesWatch
        image: fasibio/graphqldockerproxy
        name: api
        ports:
        - containerPort: 3000
        resources: {}
      restartPolicy: Always
status: {}

serice.yaml

apiVersion: v1
kind: Service
metadata:
  annotations:
    kompose.cmd: kompose convert
    kompose.version: 1.13.0 (84fa826)
  creationTimestamp: null
  labels:
    io.kompose.service: api
  name: api
  namespace: gqlproxy
spec:
  ports:
  - name: "3000"
    port: 3000
    targetPort: 3000
  selector:
    io.kompose.service: api
status:
  loadBalancer: {}

⁠The Yaml for the GraphQL (SWAPI)

Here it is importend that the service have the annotations

• gqlProxy.token: The same token you set in the proxy. (In this example 1234)
• gqlProxy.url: This is the relative path to the proxy running inside the container. (For example: :9000/graphql)
• gqlProxy.namespace: The namespace that wraps your microservice.

---
kind: Deployment
apiVersion: extensions/v1beta1
metadata:
  
  labels:
    app:     swapi
  name:      swapi
  namespace: starwars
spec:
  minReadySeconds: 20
  replicas: 2
  revisionHistoryLimit: 32
  template:
    metadata:
      name: swapi
      labels:
        app: swapi
    spec:
      terminationGracePeriodSeconds: 1
      containers:
        - image: bfillmer/graphql-swapi
          imagePullPolicy: Always
          name: swapi
          ports:
            - containerPort: 9000
              name: http-port
---
kind: Service
apiVersion: v1
metadata:
  annotations:
    gqlProxy.token: '1234'
    gqlProxy.url: ':9001/graphql'
    gqlProxy.namespace: 'swapi'
  labels:
    name: swapi
  name:      swapi
  namespace: starwars
spec:
  ports:
    - port: 9001
      targetPort: 9000
      name: http
  selector:
    app: swapi

⁠Kubernetes no Permission to scan a namespace.

If the K8s User have no Permission to Scan a namespace, this will be added to a blacklist. So this namespace will be no longer pulled. You can clear the Blacklist add the admin page.

⁠All About Namespaces

Namespaces are set by the GraphQl backend microservice, with the label gqlProxy.namespace. If you need more than one GraphQL backend server in the same namespace, then give the same name in the label gqlProxy.namespace. The proxy will merge the services.

⁠WARNING!!!!

At the moment it's not possible to have same queries, mutations or types for different entities. The proxy will use the first one it finds.

⁠Admin page / Metadata Page

To see what the proxy has included and there is another graphql service under /admin/graphql as well. Here you can see all of the namespaces and endpoint metadata for the included proxy nodes. If an endpoint being served by a loadbalancer, then you can also find the "real" endpoints.

Set the environment variables, gqlProxyAdminUser and gqlProxyAdminPassword, to configure a Basic Auth for the admin page.

⁠Available Environments for the GraphQL Proxy

⁠Possible Environment Variable Combinations for Docker

• qglProxyRuntime=dockerWatch
• dockerNetwork=web

⁠Possible Environment Variable Combinations for Kubernetes

⁠Watching:

⁠For a User in the Pod

• qglProxyRuntime=kubernetesWatch
• kubernetesConfigurationKind=getInCluster

⁠For an Explicit User

• qglProxyRuntime=kubernetesWatch
• kubernetesConfigurationKind=getInClusterByUser
• gqlProxyK8sUser=myK8sUser
• gqlProxyK8sUserPassword=thePassword

⁠Polling (the old, bad way)

⁠For a User in the Pod

• qglProxyRuntime=kubernetes

• kubernetesConfigurationKind=getInCluster

⁠For an Explicit User

• qglProxyRuntime=kubernetes
• kubernetesConfigurationKind=getInClusterByUser
• gqlProxyK8sUser=myK8sUser
• gqlProxyK8sUserPassword=thePassword

⁠For All of the Above

• gqlProxyPollingMs=10000
• gqlProxyAdminUser=myAdminPageUser
• gqlProxyAdminPassword=adminPassword

⁠Available Labels/Annotations for all GraphQL Endpoints