CI/CD with React, Docker, TravisCI, Github Actions and Netlify

This project is a simple React app that can be built locally with Docker and deployed to Netlify or AWS using TravisCI or Github Actions.

The inspiration for the workflos in this project came from Stephen Grider's Docker and Kubernetes: The Complete Guide course on Udemy. This is a great course for learning Docker and Kubernetes. I have taken his examples of deploying a React app to AWS with TravisCI and modified them to work with Netlify and GitHub Actions. I strongly encourage anyone interested in this topic to take his course.

CI/CD with React, Docker, TravisCI, Github Actions and Netlify
Fork This Repository and Pull It Locally
CI/CD with GitHub Actions to Netlify
CI/CD with TravisCI to AWS Elastic Beanstalk
CI/CD with GitHub Actions to AWS Elastic Beanstalk

Prerequisites

To follow this tutorial, you need to have the following:

Docker installed locally
A Github account (to push this code to and connect to TravisCI and Netlify, to run GitHub Actions)
A TravisCI account (to run the CI/CD pipeline and deploy to AWS)
A Netlify account (to deploy the app)
An AWS account (to create an Elastic Beanstalk instance to deploy from TravisCI and GitHub Actions)

Notes about the Project

The Dockerfile and docker-compose.yaml files are used for production. They are not used for development. The Dockerfile.dev and docker-compose-dev.yaml files are used for development - as you saw in the above steps.

Notes:

If you get an error about the port being in use, you can run docker-compose -f docker-compose-dev.yaml down to stop the containers. Or run docker stop $(docker ps -aq) && docker rm $(docker ps -aq) to stop and remove all containers.
The --build flag is only needed the first time you run the command.

Notes about the CI/CD Methods

Everything in this tutorial is free EXCEPT:

AWS Elastic Beanstalk can be free for new users, otherwise, it will cost money. Make sure you tear down the Elastic Beanstalk instance when you are done.
TravisCI offers a trial period for new users. After the trial period, you will have to pay for the service. Make sure you tear down the TravisCI pipelines when you are done.

Fork This Repository and Pull It Locally

To use this repository for any of the CI/CD methods we discussed, you can fork it to your GitHub account. Then clone it to your local machine.

git clone https://github.com/YOUR-USERNAME/ci-cd-webgeeks-feb-2023

You can see full instructions on how to do this here.

Alternatively, you could clone this repository directly to your local machine. However, you will not be able to push your changes to GitHub unless you update the remote url.

Running the App with Docker

The Dockerfile.dev looks something like this:

# Specify a base image
FROM node:18.13.0-alpine

# Set working directory to user app folder
WORKDIR '/app'

# Copy package.json and install depenendencies
COPY package.json .
RUN npm install

# Copy all the other files
COPY ./ ./

# Default command
CMD ["npm", "start"]

It is based on the node:18.13.0-alpine image. It copies the package.json file and installs the dependencies. It then copies all the other files and runs the default command npm start.

To run this Dockerfile, make sure you have Docker installed.

To build and run the app with Docker run the following commands:

docker build -f Dockerfile.dev -v -t simple-react .
docker run -p 3000:3000 simple-react

We could add volume mapping so that we don't have to rebuild the image every time we make a change to the code. To do this, we can add the following to the docker run command:

-v /app/node_modules -v $(pwd):/app

Running the Tests with Docker

While the container is running, you can get the container ID using docker ps. To run the test, run the following command:

docker exec -it <container_id> npm run test

The exec command allows you to run a command inside a running container. The -it flag allows you to run the command interactively. The npm run test command runs the tests.

You can alternatively run tests with Docker using the following command even if the container is not running.

docker run simple-react npm run test

Here npm run test is the command that is run inside the container instead of the command specified in the Dockerfile.

Running the App and Tests with Docker Compose

The docker-compose-dev.yaml file looks like this:

version: '3'
services:
  react-app:
    build:
      context: .
      dockerfile: Dockerfile.dev
    ports:
      - "3000:3000"
    volumes:
      - /app/node_modules
      - .:/app
  test:
    build:
      context: .
      dockerfile: Dockerfile.dev
    volumes:
      - /app/node_modules
      - .:/app
    command: ["npm", "run", "test"]

It specifies two services. The first service is the called react-app. It builds the image using the Dockerfile.dev file. It maps the port 3000 on the host to port 3000 on the container so we can access it in the browser. This is the same as the -p 3000:3000 flag we used in the previous step. It mounts two volumes:

The node_modules folder in the container to the node_modules folder on the host. This makes it so that we use the node_modules installed within the container (rather than the current working directory that we map next). This is the same as the -v /app/node_modules flag we used in the previous step.
The current directory on the host to the /app folder in the container. This makes it so that we don't have to rebuild the image every time we make a change to the code.

To run the app and tests with Docker Compose, run the following command:

docker-compose -f docker-compose-dev.yaml up --build

Notice that we can't interact with the tests. This is because the tests run in the background. To run the tests interactively, we can run the following command:

docker-compose -f docker-compose-dev.yaml run test

What about Dockerfile and docker-compose.yaml?

The Dockerfile looks like this:

# Specify a base image
FROM node:18.13.0-alpine as builder

# Set working directory to user app folder
WORKDIR '/app'

# Copy package.json and install depenendencies
COPY package.json .
RUN npm install

# Copy all the other files
COPY ./ ./

# Run Build
RUN npm run build

# Use nginx as the base image
FROM nginx:latest

# Copy the build directory to nginx
COPY --from=builder /app/build /usr/share/nginx/html

This file does similar steps as the Dockerfile.dev file. The main difference is that it runs the npm run build command to build the app for production. It then uses the nginx:latest image as the base image. It copies the build directory to the nginx directory. No command is specified in the Dockerfile. This is because the default command for the nginx image is to start the nginx server.

CI/CD with GitHub Actions to Netlify

Note: You do not need to use GitHub actions to deploy to Netlify. You can use Netlify's GitHub integration to deploy to Netlify directly. However, we will use GitHub actions to deploy to Netlify so you can see how to use GitHub actions, after showing you how to use Netlify's GitHub integration.

What is Netlify?

Netlify is a CI/CD tool that allows you to deploy and host your app to the web. It started with static site hosting but now supports frameworks with servers with the use of serverless functions.

You can connect GitHub repositories to Netlify. When you push code to a connected repository, Netlify will automatically build and deploy the app. We will start with this so you can see how easy it is to deploy to Netlify. Then we will add integration tests to the CI/CD pipeline using GitHub actions.

Connecting Netlify to GitHub

To connect Netlify to GitHub:

Go to Netlify Signup
Sign up using your GitHub account.
Click on the Authorize netlify button.
When given the option, choose Import an existing project -> Import from Git
Choose the repository you created
Click on the Deploy site button.

You should see a message that says Your site is building. Once the build is complete, you should see a message that says Your site is live. You can click on the View site button to see your app.

Running Tests with Netlify Build

The simplest way to get netlify to run your tests is to:

Go to Site settings -> Build & deploy -> Build settings
Click on the Edit settings button
Update the build command to:

npm run test && npm run build

This will run the tests and then build the app if the tests pass. If the tests fail, the build will fail.

To verify this, you can make a change to the code that will cause the tests to fail (or write a new failing test). Then push the code to GitHub. You should see that the build fails. You can see an example of failing tests here.

Using GitHub Actions to Run Tests and Deploy to Netlify

An example of this workflow yaml can be found on the netlify-actions branch of this repository.

While the previous method works, it is not ideal. If we had further tests we needed to run, or if we wanted to run our tests in docker, this would get complicated with the Netlify build settings. Instead, we will use GitHub actions to run our tests.

To do this, we will create a new workflow file in the .github/workflows directory. The file should be named ci.yaml. It should look like this:

name: CI

on:
  push:
    branches:
      - netlify-actions

jobs:
  build:
    runs-on: ubuntu-20.04

    steps:
      - name: Checkout code
        uses: actions/checkout@v2
      - run: docker build -t react-test -f Dockerfile.dev .
      - run: docker run -e CI=true react-test npm test
      - name: Publish
        uses: jsmrcaga/[email protected]
        with:
          NETLIFY_AUTH_TOKEN: ${{ secrets.MY_TOKEN_SECRET }}
          NETLIFY_DEPLOY_TO_PROD: true
          NETLIFY_SITE_ID: UPDATE_WITH_YOUR_SITE_ID

Before this will work, there are a few things we need to do.

Disconnect our repository from Netlify
Get the Netlify site ID
Create a Netlify access token See directions below.

Disconnect our repository from Netlify

Unlinking your repository from Netlify will make it so pushes to your repository no longer trigger deploys to github. This is because we want to use GitHub actions to deploy to Netlify.

To do this:

Go to Netlify
Go to the site you created by connecting it to the GitHub repository
Click on Site settings then Build & deploy
Click on Manage repository and then Unlink <YOUR REPOSITORY>

Get the Netlify site ID

You will need to get the Netlify site ID for our GitHub action. To do this:

Go to Netlify
Go to the site you created by connecting it to the GitHub repository
Click on Site settings then General
Copy the Site ID value
Place it in your ci.yaml file in the NETLIFY_SITE_ID field

Create a Netlify Access Token

You will need to create a Netlify access token. To do this:

Go to Netlify
Click on your profile picture in the top right corner
Click on User settings
Click on Applications
Under Personal access tokens, click on New access token
Give the token a name and click on the Create token button
IMPORTANT: Do not share this token. Copy the access key and secret access key - once you navigate away from the page AWS will not reveal the key to you again.
Copy the token and go to your GitHub repository.
Go to Settings -> Secrets and Variables -> Actions
Click on the New repository secret button
Gve the secret the name MY_TOKEN_SECRET and store the access token value here
Click on the Add secret button

Pushing Code

Commit and push and you should have a functional CI/CD pipeline testing your code before building and deploying to Netlify.

You'll notice that within the build step, GitHub actions is now building the site and pushing the deploy to Netlify. This contrasts with the previous method where Netlify was building the site and pushing the deploy. This is the desired behavior, allowing us to have more control over the build and deploy process.

CI/CD with TravisCI to AWS Elastic Beanstalk

What is TravisCI?

TravisCI is a CI/CD tool that is built into GitHub. It allows you to automate workflows. It is similar to GitHub Actions and CircleCI.

Connecting TravisCI to GitHub

To connect TravisCI to GitHub, go to travis-ci.com. Then click on the Sign in with GitHub button. Then click on the + button next to your GitHub username. Then click on the Activate button next to the repository you want to connect to TravisCI. For more details on this, check out the TravisCI documentation or this blog.

Creating a TravisCI Workflow

An example of this workflow yaml can be found on the travis-ci branch of this repository.

To create a workflow, create a .travis.yml file in the root of the project. The file should look like this:

sudo: required
services:
  - docker

before_install:
  - docker build -f Dockerfile.dev -t my_image .

script:
  - docker run -e CI=true my_image npm run test
# deploy:
#   provider: elasticbeanstalk
#   access_key_id: $AWS_ACCESS_KEY
#   secret_access_key: $AWS_SECRET_KEY
#   region: "us-east-2"
#   app: "UPDATE_WITH_YOUR_APPLICATION_NAME"
#   env: "UPDATE_WITH_YOUR_ENV_NAME"
#   bucket_name: "UPDATE_WITH_YOUR_BUCKET_NAME"
#   bucket_path: "UPDATE_WITH_YOUR_ENV_NAME"
#   on:
#     branch: main

Breaking down the file:

The sudo field specifies that we need root access to run the build. This is because we need to run docker commands.
The services field specifies the services we need to run the build. In this case, we need to run the docker service.
The before_install field specifies the commands that need to be run before the build. In this case, we need to build the image using the Dockerfile.dev file.
The script field specifies the commands that need to be run during the build. In this case, we need to run the tests in the container. This will run our tests and exit the container. If the tests fail, the build will fail.
The deploy field specifies the commands that need to be run after the build. In this case, we will deploy the app to AWS Elastic Beanstalk. This is commented out because we don't have an AWS account yet. We will uncomment this later.

Committing and pushing this code to GitHub will trigger a build on TravisCI. You can see the build status on the TravisCI website. If the build fails, you can click on the build to see the logs. You can also see the build status on the GitHub repository. The build should pass because there is only one test that runs.

Deploying to AWS Elastic Beanstalk for TravisCI

Creating Elastic Beanstock in AWS for TravisCI

To deploy to AWS Elastic Beanstalk, you will need to

Log into your AWS account
Navigate to the Elastic Beanstalk service
Create an Elastic Beanstalk application with the following:
- Your choice of name
- Platform: Docker
- Sample application
- Note: Application name will be used for UPDATE_WITH_YOUR_APPLICATION_NAME from the .travis.yml file
This will automatically create an environment for you.
- Note: Environment name will be used for UPDATE_WITH_YOUR_ENV_NAME from the .travis.yml file
Create an IAM user with the following permissions:
- AdministratorAccess-AWSElasticBeanstalk
Create an access key for the user - currently this can be found by going to:
- IAM
- Users
- Your User (the one you just created)
- Security Credentials
- Access Keys
- Create New Access Key
- On Access key best practices & alternatives, pick Third-party service, check you understand their recommendation and click continue (or figure out how to follow their direction about using IAM role instead)
- IMPORTANT: Do not share this token. Copy the access key and secret access key - once you navigate away from the page AWS will not reveal the key to you again.
Add the access key to TravisCI
Navigate to S3 and find the name of the bucket created for your Elastic Beanstalk application
- Note: Bucket name will be used for UPDATE_WITH_YOUR_BUCKET_NAME from the .travis.yml file

In TravisCI

Go to the lastest build on TravisCI. Then click on the More options button. Then click on Settings. Then add the following environment variables:

AWS_ACCESS_KEY -> from the IAM user you created noted IMPORTANT above
AWS_SECRET_KEY -> from the IAM user you created noted IMPORTANT above Make sure both leave DISPLAY VALUE IN BUILD LOG unchecked - we do not want these values to be displayed in the build log.

In travis.yml

Uncomment the deploy section in the .travis.yml file. Then update the following fields:

UPDATE_WITH_YOUR_APPLICATION_NAME -> the name of the Elastic Beanstalk application you created
UPDATE_WITH_YOUR_ENV_NAME -> the name of the Elastic Beanstalk environment you created
UPDATE_WITH_YOUR_BUCKET_NAME -> the name of the directory within the S3 bucket that is generated for your Elastic Beanstalk application. It should contain the word elasticbeanstalk and the region you are using.
Note: Update region if your region is not us-east-2

Push your code to GitHub. This will trigger a build on TravisCI. The build should pass and the app should be deployed to AWS Elastic Beanstalk.

CI/CD with GitHub Actions to AWS Elastic Beanstalk

What is GitHub Actions?

GitHub Actions are a CI/CD tool that is built into GitHub. It allows you to automate workflows. It is similar to TravisCI and CircleCI.

Creating a Github Actions Workflow

An example of this workflow yaml can be found on the github-actions branch of this repository.

To create a workflow, create a .github/workflows folder in the root of the project. Then create a file called ci.yml in the folder. The file should look like this:

name: CI

on:
  push:
    branches:
      - main

# env:
#   AWS_ACCESS_KEY: ${{ secrets.AWS_ACCESS_KEY }}
#   AWS_SECRET_KEY: ${{ secrets.AWS_SECRET_KEY }}

jobs:
  build:
    runs-on: ubuntu-20.04

    steps:
      - name: Checkout code
        uses: actions/checkout@v2
      - run: docker build -t react-test -f Dockerfile.dev .
      - run: docker run -e CI=true react-test npm test

      # - name: Generate deployment package
      #   run: zip -r deploy.zip . -x '*.git*'

      # - name: Deploy to EB
      #   uses: einaregilsson/beanstalk-deploy@v21
      #   with:
      #     aws_access_key: ${{ env.AWS_ACCESS_KEY }}
      #     aws_secret_key: ${{ env.AWS_SECRET_KEY }}
      #     region: us-east-2
      #     application_name: UPDATE_WITH_YOUR_APPLICATION_NAME
      #     environment_name: UPDATE_WITH_YOUR_ENV_NAME
      #     existing_bucket_name: UPDATE_WITH_YOUR_BUCKET_NAME
      #     version_label: ${{ github.sha }}
      #     deployment_package: deploy.zip
      #     use_existing_version_if_available: true