Coding for Entrepreneurs

CI/CD Github Workflow for Django & DigitalOcean App Platform

By Justin

CI/CD Github Workflow for Django & DigitalOcean App Platform
Whenever you aim to release code into production, it's best that you follow pre-defined steps to ensure your application is working as intended.
Perhaps the most important step is to run your automated tests. Below is a github workflow to help you do exactly that.
After we learn to run a github workflow to test our Django code, we'll implement a step to trigger an update on our production application.
Let's keep in mind that this is meant to be a basic working example. You can, and probably should, look into automating more tasks (such as linting, coverage reports, collecting static files, building JS apps, etc) using github workflows.

Requirements

1. Store your DigitalOcean API Token

For Github Actions, you need to add the following Repository Secrets (located in your repo's settings under settings/secrets/actions):
  • PERSONAL_ACCESS_TOKEN this is your Github Personal Access token. Create one here.
  • DO_ACCESS_TOKEN this is your DigitalOcean access token created here.
  • DO_APP_ID: Your App Platform id -- Get this with doctl apps list --format "Spec.Name, ID" What's doctl? Check out this guide.

2. Setup the continuous integration (CI) for Django using Postgres & Running Tests

.github/workflows/main.yaml
yaml
# name of our workflow
name: Django CI/CD Workflow

# triggers for our workflow
on:
  push:
    branches: [ main ]
  pull_request:
    branches: [ main ]

jobs:
  build:
    runs-on: ubuntu-latest
    services:
      postgres_main:
        image: postgres:12
        env:
          POSTGRES_USER: postgres
          POSTGRES_PASSWORD: postgres
          POSTGRES_DB: djtesting
        ports:
          - 5432:5432
        options: --health-cmd pg_isready --health-interval 10s --health-timeout 5s --health-retries 5
    steps:
        - name: Checkout code
          uses: actions/checkout@v2
        - name: Set up Python 3.6
          uses: actions/setup-python@v2
          with:
              python-version: 3.6
        - name: Install requirements
          run: pip install -r requirements.txt
        - name: Run Tests
          env:
            DJANGO_SECRET_KEY: CI_CD_TEST_KEY
            POSTGRES_DB: djtesting
            POSTGRES_HOST: localhost
            POSTGRES_USER: postgres
            POSTGRES_PASSWORD: postgres
            POSTGRES_PORT: 5432
            DEBUG: "0"
          run: |
            python manage.py test
One of the biggest keys here is that the POSTGRES_HOST is set to localhost and not the workflow service declaration postgres_main. Much of the service configuration emulates docker-compose but it's not an exact match while using.

3. Push Code to a new Branch.

Append this to the end of your steps above
yaml
        - name: Push main branch into production
          uses: codingforentrepreneurs/action-branch-to-branch@main
          with:
              dest_branch: 'production-3'
              source_branch: 'main'
              commit_message: 'Release production version'
          env:
              GITHUB_TOKEN: ${{ secrets.PERSONAL_ACCESS_TOKEN }}
The step above will override your destination branch (dest_branch) and uses this action. I only do this as a way to decide which branch I want digitalocean to run on (assuming it's different than main).

4. Add the deployment step (cd) for DigitalOcean doctl

In this post we show you the basics of using doctl and an App Specification file for DigitalOcean services. In our case here, we're using App Platform but you can adapt below (your your app.yaml file) for any of DigitalOcean's services.
Append this to the end of your steps above
yaml
        - name: Install doctl
          uses: digitalocean/action-doctl@v2
          with:
            token: ${{ secrets.DO_ACCESS_TOKEN }}
        - name: Update DigitalOcean App Platform
          run: doctl apps update ${{ secrets.DO_APP_ID }} --spec app.yaml
Discover Posts