This repository has been archived by the owner. It is now read-only.
Fetching contributors…
Cannot retrieve contributors at this time
187 lines (140 sloc) 6.19 KB

Deprecation Notice

This project is no longer being maintained by the Kelda team. To get updated when our next project is ready, sign up for our mailing list!

Build Status Go Report Card Code Coverage


A straightforward way to configure containerized applications.

What is Kelda?

Kelda is a simple way to configure, run, and manage containerized applications in the cloud.

No more YAML, templating, complicated CLIs, and new domain specific languages. Kelda’s intuitive JavaScript API makes it easy to describe what you want to run in the cloud — that is, the virtual machines, containers, network policies, and more.

Using Kelda's CLI, simply pass your high-level blueprint to Kelda’s Kubernetes-based deployment platform. We'll take care of the nitty gritty details needed to get your system up and running in your cloud of choice.


Install Kelda with npm:

$ npm install -g @kelda/install

Check out more in our Quick Start tutorial.

Deploy Quickly On...


How Do I Use It?

  1. Describe your application using Kelda's JavaScript API (or use a pre-made blueprint).

    const kelda = require('kelda');
    // A Docker container running your application.
    const app = new kelda.Container({
      name: 'my-app',
      image: 'myDockerImage',
    // Expose the application to the public internet on port 80.
    kelda.allowTraffic(kelda.publicInternet, app, 80);
  2. Describe where to run the application.

    // A virtual machine with 8 GiB RAM and 2 CPUs on Amazon EC2.
    const virtualMachine = new kelda.Machine({
      provider: 'Amazon', // 'Google', 'DigitalOcean'.
      ram: 8,
      cpu: 2,
    // An infrastructure with one Kelda Master machine and 6 Worker machines -- all in EC2.
    const infrastructure = new kelda.Infrastructure({
      masters: virtualMachine,
      workers: virtualMachine.replicate(6),
    // Deploy the app to the infrastructure.
  3. Boot the application to your preferred cloud using Kelda's CLI.

    $ kelda run ./myBlueprint.js
    Your blueprint is being deployed. Check its status with `kelda show`.
  4. Manage your deployment from the command line.

    Keep track of virtual machines and containers.

    $ kelda show
    MACHINE         ROLE      PROVIDER    REGION       SIZE         PUBLIC IP         STATUS
    i-0d1d363a57    Master    Amazon      us-west-1    m3.medium      connected
    i-0b6ef239c7    Worker    Amazon      us-west-1    m3.medium      connected
    bce2362f3a1f    i-0b6ef239c7                     my-app      running    20 seconds ago

    SSH into containers and machines (here, the my-app container).

    $ kelda ssh my-app

    Stop the deployment.

    $ kelda stop

    And more...

Learn more in our Quick Start tutorial.


Here are a some of the cool things you can do with Kelda's API.

Share and Reuse Configuration

Don't waste time combing through reams of application specific documentation. Share and reuse configuration like you do with other software libraries.

const Redis = require('@kelda/redis');
const redisDB = new Redis(3, 'ADMIN_PASSWORD'); // A Redis database with 3 workers.

Configure a Secure Network

Only allow the traffic your application needs.

// Let the app send traffic to the Redis database on port 6379.
kelda.allowTraffic(app, redisDB.master, 6379);

Run Containers

Quickly get running in the cloud.

const otherApp = new kelda.Container({
  name: 'my-other-app',
  image: 'myOtherImage',

Keep Sensitive Data Safe

We keep your keys, certificates, and passwords encrypted and in the right hands.

otherApp.env['GITHUB_OATH_TOKEN'] = new kelda.Secret('githubOathToken');

Deploy Anywhere

Kelda's API is cloud agnostic, so you, your users, and colleagues can run the same blueprint in any of Kelda's supported clouds.

const machine = new kelda.Machine({
  provider: 'Google', // 'Amazon', 'DigitalOcean'.
  size: 'n1-standard-2', // 'm3.medium', '1gb'.

See the full API reference in our docs.

Reference Blueprints

Kelda blueprints have been written for many popular applications. Here are some that we recommend checking out:

  • Apache Spark. Deploy a multi-node Spark cluster for data processing.
  • HAProxy. Load balance traffic across different applications.
  • Kibana. Vizualize data from a hosted Elasticsearch instance.

Get Involved

We would love to work with you, get your feedback, or answer any questions you have!