Project Structure

...

• We use Docker images to deploy the applications
• I’m assuming everyone is familiar with Docker but if not let me know and I’ll run through why we use it
• Docker images can have tags, so we use these to separate environments (dev, staging, production etc)
• Our Docker-Compose file pulls and configures all images required to run the whole stack
• Runtime configuration is handled by .env files or environment variables

Devops Flow

Branches

Image Removed

• Main
• Development branchCurrent "snapshot" branch
• Code here should always run (if not stable)
• Code should rarely never be committed directly
• Base for feature/fix branches
• Feature branches
• When working on a specific feature/fix, create a new branch for that work item 
• Either submit a PR when ready, or create a draft PR whenever
• PRs should pass all tests and code-quality checks before merging
• PRs should be reviewed by someone else on the project before merging
• Staging
• Once the “main” branch is ready to be deployed to staging, merge into the ‘staging’ branch
• This will trigger an image build with the “staging” tag
• Production
• Once the “staging” branch is ready to be deployed to live, merge into the ‘production’ branch
• This will trigger an image build with the “production” tag

GitHub Actions

• All testing and build automation is handled by GitHub Actions
• If you’ve used Bamboo it’s the same thing but with a sensible configuration system
• Example config file:
• https://github.com/renalreg/ukrdc-fastapi/blob/main/.github/workflows/main.yml

How to deploy a new release

Build and publish Docker images

This guide is broadly applicable to both "ukrdc-fastapi" and "ukrdc-nuxt"

• Navigate to Actions

Image Added

• From the Publish Docker Image workflow, click Run Workflow

Image Added

• Enter the deployment environment you wish to deploy to
  • This will become the docker image tag
  • "production" and "staging" are protected, and will wait until the deployment has been approved by a project admin

Image Added

Update containers on the app server

• Once the image has been published, connect to the app server via SSH
• Run:

• cd /srv/ukrdc-compose

• docker-compose pull 

• docker-compose down

• docker-compose up -d

• Ideally also run

• docker system prune

• You can view real-time logs from the entire stack with:

• docker-compose log -f

Active Instances

• staging-app and live-app both include new UKRDC instances
• /srv/ukrdc-compose/
• In both cases, the host nginx config needs to be set up to forward to the Docker stack
• E.g. /etc/nginx/conf.d/staging-app.ukrdc.nhs.uk.conf
• This should not need to be changed unless we want to change the new stack base URL

...

• Our Docker images are deployed to GHCR (https://github.com/orgs/renalreg/packages)
• Once the image has been deployed (check repo Actions tab for status), connect to the app server via SSH
• Run:
• cd /srv/ukrdc-compose
• docker-compose pull 
• docker-compose down
• docker-compose up -d

...