Nodejs Docker And Aws

Jul 31, 2017

3 minute read

nodejs

express

docker

Elasticsearch

In this post, we will make use of NodeJS and Docker to provide and API which interfaces with an Elasticsearch instance in AWS.

Assuming you already have an AWS account, and aws-cli set up.

Process

Write the NodeJS app locally
Wrap it a docker image and deploy this image to a registry (we will use docker hub here)
Use AWS ECS (EC2 Container Service) to deploy and scale your app in the cloud.

Write the NodeJS app locally

If you have your crendentials in ~/.aws/credentials, the connection to your Elasticsearch instance will be picked up automatically. No need to worry about that.

We’re using here Jetbrain’s Webstorm and a new Express project comes with several dependencies, some of which we don’t need. Still, let’s keep them for simplicity and add some others related to the project, in particular elasticsearch, http-aws-es and aws-sdk.

package.json

{
  "name": "gardenlog-nodejs",
  "version": "0.0.0",
  "private": true,
  "scripts": {
    "start": "node ./bin/www"
  },
  "dependencies": {
    "aws-sdk": "^2.93.0",
    "body-parser": "~1.17.1",
    "cookie-parser": "~1.4.3",
    "debug": "~2.6.3",
    "elasticsearch": "^13.2.0",
    "express": "~4.15.2",
    "http-aws-es": "^2.0.5",
    "jade": "~1.11.0",
    "morgan": "~1.8.1",
    "serve-favicon": "~2.4.2"
  }
}

Docker container

With Webstorm you can use docker as your remote node interpreter. Please refer to the relevant documentation for use.

Just add a Dockerfile at the root of your project.

Dockerfile

FROM node:boron

# Create app directory
WORKDIR /usr/src/app

# Install app dependencies
COPY package.json .
# For npm@5 or later, copy package-lock.json as well
# COPY package.json package-lock.json .

RUN npm install

# Bundle app source
COPY . .

EXPOSE 3000
CMD [ "npm", "start" ]

You are now ready to run your app in a container.

Run your container locally

aws cli gets credentials in a specific order and so far, we’ve used the ones stored in the credentials file. This is fine because your local machine is where the cli is installed. But the docker container has no idea about this file. We could share it or copy it with the container. Another option, more production-friendly is to pass credentials as environment variables. It’s described below.

With docker cli commands

Build your image

docker build -t <your username>/node-web-app .

Grab your credentials, we will pass them as environment variables.

cat ~/.aws/credentials

[default]
aws_access_key_id = YOUR_KEY_ID
aws_secret_access_key = YOUR_SECRET

Run it.

docker run -p 49160:3000 -d -e "AWS_ACCESS_KEY_ID=YOUR_KEY_ID" -e "AWS_SECRET_ACCESS_KEY=YOUR_SECRET" <your username>/node-web-app

You should now see your app at http://localhost:49160
When you’re happy with it, push it to your registry
```
docker push <your username>/node-web-app
```

If you’re not happy and need to amend your code, you need to repeat all these steps.

Configure Webstorm for it

Webstorm allows us to go far beyond that, and particularly use the debugger with Docker for instance. We can also teach him where to push images, and many other features.

There are two things to consider: