This tutorial explains the advantages of working Node.js purposes in Docker containers and the best way to create a sensible improvement workflow.

Node.js permits you to create quick and scalable net apps utilizing JavaScript on the server in addition to on the consumer. Your app might run completely in your improvement machine, however are you able to make sure it’ll run in your colleague’s gadgets or manufacturing servers?

Take into account these eventualities:

  • You could be utilizing macOS when others use Home windows and the server runs Linux.
  • You could have Node.js 20 put in, however others use a variety of runtime variations.
  • You’re utilizing dependencies corresponding to databases, which have variations or will not be out there on different platforms.
  • Are you positive your new code can’t do something harmful on one other working system (OS)?
Desk of Contents

Docker Delivers

Docker helps to resolve these “however it works on my machine” points listed above. Reasonably than putting in an software domestically, you run it in a light-weight remoted digital machine-like setting referred to as a container.


An actual digital machine emulates PC {hardware} so you may set up an OS. Docker emulates an OS so you may set up purposes. It’s typical to put in one app per Linux-based container and join them by way of a digital community to allow them to talk on HTTP ports.

The benefits:

  • Your Docker setup can both emulate a manufacturing Linux server or you may deploy utilizing containers.
  • You may obtain, set up, and configure dependencies in minutes.
  • Your containerized app runs identically throughout all gadgets.
  • It’s safer. Your app may trash a container’s OS, however it received’t have an effect on your PC and you may restart afresh in seconds.

With Docker, there’s no want to put in Node.js in your PC or use a runtime administration choice corresponding to nvm.

Your First Script

Set up Docker Desktop on Windows, macOS, or Linux then create a small script named model.js with the next code:

console.log(`Node.js model: ${ course of.model }`);

When you’ve got Node.js put in domestically, attempt working the script. You’ll see the output corresponding to this in the event you had model 18 put in:

$ node model.js
Node.js model: v18.18.2

Now you can run the identical script inside a Docker container. The command under makes use of the latest long-term help (LTS) model of Node.js. cd into the script’s listing and run it on macOS or Linux:

$ docker run --rm --name model 
  -v $PWD:/residence/node/app 
  -w /residence/node/app 
  node:lts-alpine model.js

Node.js model: v20.9.0

Home windows Powershell customers can use an identical command with {} brackets round PWD:

> docker run --rm --name model -v ${PWD}:/residence/node/app -w /residence/node/app node:lts-alpine model.js

Node.js model: v20.9.0

The primary run might take a minute or two to execute as Docker downloads dependencies. Subsequent runs are instantaneous.

Let’s attempt a special model of Node — corresponding to the most recent launch of model 21. On macOS or Linux:

$ docker run --rm --name model 
  -v $PWD:/residence/node/app 
  -w /residence/node/app 
  node:21-alpine model.js

Node.js model: v21.1.0

On Home windows Powershell:

> docker run --rm --name model -v ${PWD}:/residence/node/app -w /residence/node/app node:21-alpine model.js

Node.js model: v21.1.0

Bear in mind the script is working inside a Linux container which has a particular model of Node.js put in.

Argument rationalization

For the curious, the command arguments are:

  • docker run begins a brand new container from an picture — extra about that under.

  • --rm removes the container when it terminates. It’s not essential to retain containers except you will have good cause to restart them once more.

  • --name model assigns a reputation to the container for easier administration.

  • -v $PWD:/residence/node/app (or -v ${PWD}:/residence/node/app) bind mounts a quantity. On this case, the present instantly on the host PC is mounted contained in the container at /residence/node/app.

  • -w /residence/node/app units the Node.js working listing.

  • node:lts-alpine is the picture — on this case, the LTS model of Node.js working in Alpine Linux. The picture accommodates the OS and recordsdata required to run an software. Consider it as a disk snapshot. You can begin any variety of containers from the identical picture: all of them reference the identical set of recordsdata so every container requires minimal assets.

  • model.js is the command to execute (from contained in the working listing).

Docker photos can be found from Docker Hub and so they’re out there for purposes and runtimes together with Node.js. Photographs are sometimes out there in a number of variations recognized with a tag corresponding to :lts-alpine, 20-bullseye-slim, or simply newest.

Word that Alpine is a tiny Linux distribution with a base picture dimension of round 5MB. It doesn’t comprise many libraries, however it’s ok for easy initiatives corresponding to these on this tutorial.

Operating Advanced Purposes

The model.js script above is straightforward and accommodates no dependencies or construct steps. Most Node.js purposes use npm to put in and handle modules in a node_modules listing. You may’t use the command above as a result of:

  • You may’t run npm on the host PC (it’s possible you’ll not have the Node.js or the right model put in).
  • Some modules require platform particular binaries. You may’t set up a Home windows binary on the host PC and count on it to run in a Linux container.

The answer is to create your personal Docker picture containing:

  • an applicable model of the Node.js runtime
  • an put in model of your app with all required modules

The next demonstration builds a easy Node.js app utilizing the Express.js framework. Create a brand new listing named easy and add a bundle.json file with the next content material:

  "title": "easy",
  "model": "1.0.0",
  "description": "easy Node.js and Docker instance",
  "sort": "module",
  "major": "index.js",
  "scripts": {
    "debug": "node --watch --inspect= index.js",
    "begin": "node index.js"
  "license": "MIT",
  "dependencies": {
    "specific": "^4.18.2"

Add an index.js file with JavaScript code:

import specific from 'specific';

const cfg = ;

const app = specific();

app.get('/:title?', (req, res) => {
  res.ship(`Hi there $!`);

app.hear(cfg.port, () => {
  console.log(`server listening at http://localhost:${ cfg.port }`);

Don’t try to put in dependencies or run this app on the host PC!

Create a file named Dockerfile with the next content material:

# base Node.js LTS picture
FROM node:lts-alpine

# outline setting variables
ENV HOME=/residence/node/app
ENV NODE_ENV=manufacturing

# create software folder and assign rights to the node person
RUN mkdir -p $HOME && chown -R node:node $HOME

# set the working listing

# set the lively person
USER node

# copy bundle.json from the host
COPY --chown=node:node bundle.json $HOME/

# set up software modules
RUN npm set up && npm cache clear --pressure

# copy remaining recordsdata
COPY --chown=node:node . .

# expose port on the host

# software launch command
CMD [ "node", "./index.js" ]

This defines the steps required to put in and execute your app. Word that bundle.json is copied to the picture, then npm set up is run earlier than copying the remaining recordsdata. That is extra environment friendly than copying all recordsdata directly, as a result of Docker creates a picture layer at each command. In case your software recordsdata (index.js) change, Docker want solely run the ultimate three steps; it doesn’t have to npm set up once more.

Optionally, you may add a .dockerignore file. It’s just like .gitignore and stops pointless recordsdata being copied into the picture by COPY . .. For instance:




Construct a Docker picture named easy by coming into the next command (word the . interval on the finish — which denotes you’re utilizing recordsdata within the present listing):

$ docker picture construct -t easy .

The picture ought to construct inside a couple of seconds if the node:lts-alpine Docker picture used above hasn’t been deleted out of your system.

Assuming the construct is profitable, begin a container out of your picture:

$ docker run -it --rm --name easy -p 3000:3000 easy

server listening at http://localhost:3000

The -p 3000:3000 publishes or exposes a <host-port> to a <container-port> so port 3000 in your host PC routes to port 3000 contained in the container.

Open a browser and enter the URL http://localhost:3000/ to see “Hi there World!”

Attempt including names to the URL — corresponding to http://localhost:3000/Craig — to see various messages.

Lastly, cease your app working by clicking the cease icon within the Containers tab of Docker Desktop, or enter the next command in one other terminal window:

docker container cease easy

A Higher Docker Growth Workflow

The method above has some irritating flaws:

  • Any change to your code (in index.js) requires you to cease the container, rebuild the picture, restart the container, and retest.

  • You may’t connect a Node.js debugger such because the one out there in VS Code.

Docker can enhance your improvement workflow by retaining the prevailing, production-level picture, however working a container with overrides with a view to do the next:

  • Set setting variables corresponding to NODE_ENV to improvement.

  • Mount the native listing into the container.

  • Begin the app with npm run debug. This runs node --watch --inspect= index.js, which restarts the app when recordsdata change (new in Node.js 18) and begins the debugger with requests permitted from exterior the container.

  • Exposes app port 3000 and debugger port 9229 to the host.

You are able to do this with one lengthy docker run command, however I want to make use of Docker Compose. It’s put in with Docker Desktop and is usually used to start out multiple container. Create a brand new file named docker-compse.yml with the next content material:

model: '3'


      - NODE_ENV=improvement
      context: ./
      dockerfile: Dockerfile
    container_name: easy
      - ./:/residence/node/app
      - "3000:3000"
      - "9229:9229"
    command: /bin/sh -c 'npm set up && npm run debug'

Begin your app working in debug mode with:

$ docker compose up

[+] Constructing 0.0s
[+] Operating 2/2
 ✔ Community simple_default  Created
 ✔ Container easy        Created
Attaching to easy
easy  |
easy  | updated, audited 63 packages in 481ms
easy  |
easy  | > [email protected] debug
easy  | > node --watch --inspect= index.js
easy  |
easy  | Debugger listening on ws://
easy  | For assist, see:
easy  | server listening at http://localhost:3000

Word that older variations of Docker Compose are Python scripts run utilizing docker-compose. Newer variations have Compose performance built-in into the principle executable, so it’s run with docker compose.

Reside software restarts

Open index.js, make a change (such because the string on line 14), and save the file to see the appliance mechanically restart:

easy  | Restarting 'index.js'
easy  | Debugger listening on ws://
easy  | For assist, see:
easy  | server listening at http://localhost:3000

Open or refresh your browser at https://localhost:3000/ to view the replace.

Debug with VS Code

Open the VS Code Run and Debug panel and click on create a launch.json file.

VS Code Run and Debug pane

Select Node.js within the dropdown and a .vscode/launch.json file is created and opened within the editor. Add the next code which attaches the debugger to the working container:

  "model": "0.2.0",
  "configurations": [
      "type": "node",
      "request": "attach",
      "name": "Attach to Container",
      "address": "localhost",
      "port": 9229,
      "localRoot": "${workspaceFolder}",
      "remoteRoot": "/home/node/app",
      "skipFiles": [

Save the file then click on Connect to Container on the high of the Debug pane to start out debugging.

VS Code Run and Debug

A debugging toolbar seems. Swap to index.js and add a breakpoint to line 14 by clicking the gutter to indicate a crimson dot.

set breakpoint in VS Code

Refresh https://localhost:3000/ in your browser and VS Code will halt execution on the breakpoint and present the state of all software variables. Click on an icon on the debugging toolbar to proceed working, step by means of the code, or disconnect the debugger.

Cease the container

Cease the working container by opening one other terminal. cd to the appliance listing, and enter:

docker compose down


Whereas Docker requires some preliminary set-up time, the long-term advantages of strong, distributable code greater than outweigh the hassle. Docker turns into invaluable while you add additional dependencies corresponding to databases.

This tutorial explains the fundamentals of working Node.js apps in Docker containers. To delve additional, think about these Pylogix assets: