GithubHelp home page GithubHelp logo

actions-container-helpers's Introduction

Helpers for GitHub Actions and Docker Containers

This repository contains a set of helper scripts for working with containers in GitHub Actions.

Creating a container image

You can use the container-image.yml workflow to create a new container image. This workflow makes sure that the CI machine is setup for multi-arch, such that you can build a docker container for multiple architectures. It also logs in to the GitHub container registry allowing you to upload an image right away. To use it, add a job to your GitHub workflow calling this workflow:

# ...

jobs:

  # ...

  docker:
    uses: "tweedegolf/actions-helpers/.github/workflows/container-image.yml@main"
    with:
        push: ${{ github.ref == 'refs/heads/main' }}
        platforms: "linux/amd64,linux/arm64"
        tags: ghcr.io/tweedegolf/example:latest

  # ...

Workflow inputs

This workflow has several parameters that allow customizing the behavior:

OptionRequiredDefault
tags yes
List of image tags (newline separated) for the resulting container image
push no false
If true, the image will be pushed to the registry
platforms no ""
Comma separated list of platforms to build the image for, i.e. linux/amd64,linux/arm64. If left empty, will only build for the native platform.
context no .
Context directory for the container image build
build-args no ""
List of build arguments (newline separated) to be inserted in the container image build
file no Dockerfile
Name of the dockerfile to build
artifact no
Download artifact with this name

Example usage

Below, you will find a working example where we either just build (and not push) a container image on a pull request, and we fully build and push a container on the main branch, these two require separate permissions. We'll also add a build matrix to allow multiple images to be generated:

  • .github/workflows/docker.yml

    name: Docker
    
    on:
      workflow_call:
    
    jobs:
      docker:
        strategy:
          matrix:
            include:
              - version: trixie
                latest: false
                alt: testing
              - version: bookworm
                latest: true
                alt: stable
              - version: bullseye
                latest: false
                alt: oldstable
        uses: "tweedegolf/actions-helpers/.github/workflows/container-image.yml@main"
        with:
            push: ${{ github.ref == 'refs/heads/main' }}
            platforms: "linux/amd64,linux/arm64"
            build-args: |
                DEBIAN_VERSION=${{matrix.version}}
            tags: |
                ghcr.io/tweedegolf/debian:${{matrix.version}}
                ghcr.io/tweedegolf/debian:${{matrix.alt}}
                ${{ matrix.latest && 'ghcr.io/tweedegolf/debian:latest' || '' }}

    This file will be re-used by the two workflows below. As such we only trigger it on workflow_call. Note how we use a matrix to build multiple images. Of course for this workflow to run we'll also need a Dockerfile to be built, but that has been omitted in this example.

  • .github/workflows/build-push.yml

    name: Build and push
    
    permissions:
      contents: read
      packages: write
    
    on:
      push:
        branches:
          - main
      schedule:
        - cron: '30 2 * * SUN'
    
    jobs:
      build-and-push:
        uses: ./.github/workflows/docker.yml

    This build and push workflow for the main branch also runs on a schedule every week to keep the image up to date. Note how we require package write permission in this workflow.

  • .github/workflows/check.yml

    name: Checks
    
    permissions:
      contents: read
    
    on:
      pull_request:
    
    jobs:
      build:
        uses: ./.github/workflows/docker.yml
        secrets: inherit

    The checks workflow will just need read permissions for the repository with no write permissions required. We only run it for a pull request.

Cleanup of old data in the container registry

The GitHub container registry does not automatically remove old tags or untagged images from itself. To do this, we can setup an actions workflow that does this for us automatically. There are two steps to this process. First, we remove any tags that are no longer needed by adding any number of jobs refering to the container-tag-cleanup.yml workflow. Finally we finish by adding a single container-untagged-cleanup.yml workflow. Of course, if your image creation workflow never creates any new tags but instead always reuses the existing ones there is no need to include any of the tag cleanup workflows. Let's look at an example:

name: Container Registry Cleanup

permissions:
  contents: read
  packages: write

on:
  workflow_dispatch:
  schedule:
    - cron: '30 2 * * MON'

jobs:
  old-nightly-cleanup:
    uses: "tweedegolf/actions-helpers/.github/workflows/container-tag-cleanup.yml@main"
    with:
      package: debian
      filter: "^nightly-\\d{2}-\\d{2}-\\d{4}$"
      keep_n: 5
  untagged-cleanup:
    needs: [old-nightly-cleanup]
    uses: "tweedegolf/actions-helpers/.github/workflows/container-untagged-cleanup.yml@main"
    with:
        package: debian

In this example, we run a cleanup job that removes old nightly tagged images first, and then removes any remaining untagged images after that. Note how this action is only run on a schedule once a week, there is no need to run it after every main branch update or every pull request unless lots of container builds are happening on your repository. You might have additional jobs that run the tag cleanup workflow if there are multiple patterns you wish to match, but we always make sure that the untagged cleanup runs after all other jobs have finished to make sure we capture any additional untagged images from the other cleanup jobs.

Tag cleanup workflow inputs

This workflow has several parameters that allow customizing the behavior:

OptionRequiredDefault
package yes
Name of the container registry package
filter no ""
A regular expression matching some specific tags for the container
keep_filter no ""
A regular expression matching any tags that should not be removed
keep_n no -1
How many of the most recently created and matched tags to keep (if non-negative).
older_than no -1
Only remove untagged container images at least as old as this (in days), only works if larger than zero.

Untagged images cleanup workflow inputs

This workflow has several parameters that allow customizing the behavior:

OptionRequiredDefault
package yes
Name of the container registry package
older_than no -1
Only remove untagged container images at least as old as this (in days), only works if larger than zero.
untagged_timestamp_tolerance no 10000
Do not remove untagged container images if they have a created timestamp this close (in milliseconds) to a tagged container image. We do this because multi-arch docker containers will be pushed as separate untagged images to the container registry and GitHub cannot recognize them as part of a tagged image. Setting this option large enough prevents accidental deletion of one of these images.

actions-container-helpers's People

Contributors

rnijveld avatar

Watchers

abudaan avatar Marlon avatar Erik Jonkers avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.