AlexsJones / apollo-server

Apollo server on k8s with helm example

Apollo Server

A example example of Apollo server run on kubernetes with helm.

Developed locally with skaffold

Apollo server resolvers

The server example has a resolvers.js which builds with the docker image and works to show us a basic example.

We can inject a new resolvers file through a configmap via k8s to power up the server with additional resolvers as shown templates/configmap-resolvers.yaml

kubectl exec test-apollo-8488d5b6f6-h9xll cat resolvers.js
const movies = [
  {
    title: 'Star wars',
    releaseDate: '1977'
  },
  {
    title: 'Jurassic Park',
    releaseDate: '1993'
  },
];

const resolvers = {
  Query: {
    movies: () => movies,
  },
};

module.exports = resolvers;

Here we have deployed in our configmap of resolvers, this is a nice solution when combined with versioned helm releases

Run locally before injection:

After:

Skaffold

For local reloading whilst working within server/*.js just run skaffold dev from the root directory (assuming k8s and minikube installed)

node_modules

This kind of repo structure for educational purposes causes some performance issues with helm. Large node_modules folders will stop helm gzip and configmap deploying so they need to be added to .helmignore

Helm Deployment

helm install . --name=apollo-server-release -n apollo

Helm Undeployment

helm remove apoll-server-release --purge