Skip to content

Docker Pipelines with GitHub Actions

Overview

Author a GitHub Actions workflow that builds a multi-stage Dockerfile with Buildx, produces multi-architecture images, and pushes immutable SHA tags to GitHub Container Registry (GHCR) — ready for later Kubernetes or cloud promotion.

CI builds containers so every merge produces a reproducible Open Container Initiative (OCI) image. Prefer Docker Buildx over ad-hoc docker build on a laptop. Tag with the commit SHA (and optionally a digest); promote that same image through staging and production. GHCR (ghcr.io/<owner>/<image>) is the natural home for GitHub-native pipelines; the same pattern works for Docker Hub, Amazon Elastic Container Registry (ECR), and others with different login steps.

This is a core tutorial in Module 7 · Docker Pipelines of the REBASH Academy GitHub Actions for Cloud & DevOps Engineers series — written for Cloud, DevOps, Platform, and SRE engineers.

Prerequisites

Learning Objectives

By the end of this tutorial, you will be able to:

  • Sketch a Buildx workflow with packages: write for GHCR
  • Write a minimal multi-stage Dockerfile
  • Build and push linux/amd64 and linux/arm64
  • Explain SHA tags vs floating latest
  • Describe image promotion without rebuild

Architecture

This topic’s control points and relationships are shown below.

Docker build pipeline

Theory

What it is

A Docker pipeline compiles application code into an OCI image inside a GitHub Actions job, then pushes it to a registry. Production jobs use Buildx (BuildKit) for layer caching, multi-platform manifests, and provenance-friendly builds. Multi-stage Dockerfiles keep compilers and test tools out of the final runtime image. Authentication to GHCR uses the job’s GITHUB_TOKEN with packages: write; cloud registries usually prefer OpenID Connect (OIDC) over long-lived passwords (Modules 5 and 10).

Concern Prefer Avoid
Builder Buildx + BuildKit Legacy builder only
Tags Commit SHA / semver / digest Only latest
Platforms Explicit linux/amd64,linux/arm64 Implicit host arch only
Auth GITHUB_TOKEN / OIDC Personal access tokens in Git

Why it matters

Laptop-built images drift from CI and skip scanners. Registry-hosted, SHA-tagged images are the unit of deploy for Kubernetes and managed container services. Multi-arch manifests stop “works on my amd64 runner, fails on arm64 nodes.” Without promotion discipline — retag or deploy the same digest — staging and production silently diverge when each environment rebuilds from source.

How it works

  1. Workflow checks out the repository and sets up Buildx (docker/setup-buildx-action).
  2. Log in to GHCR with docker/login-action and GITHUB_TOKEN (or OIDC for cloud registries).
  3. docker/build-push-action builds the Dockerfile, often with GitHub Actions cache (type=gha) or a registry cache.
  4. Tags include ghcr.io/<owner>/<name>:<sha> and optionally a branch or semver tag; record the digest as a job output.
  5. Deploy jobs (Modules 8–10) pull that SHA or digest; production never rebuilds from source for the same commit.

Pin base images by digest in production Dockerfiles. Cache accelerates rebuilds; it does not replace reproducible tags.

Key concepts and comparisons

Practice Good Risky
Promote Same digest across envs Rebuild per environment
Permissions Least-privilege packages: write Broad write-all
Secrets Build-args for non-secrets only Baking API keys into layers
Actions Pin by commit SHA Floating @v1 / @main

Common pitfalls

  • Pushing latest from every pull request.
  • Privileged Docker-in-Docker on shared self-hosted runners without isolation.
  • Assuming cache guarantees bit-identical images across builders.
  • Forgetting packages: write / package visibility so GHCR push fails.
  • Rebuilding for production instead of promoting the tested digest.

Hands-on Lab

Create a workspace for this tutorial.

mkdir -p ~/rebash-github-actions/docker-lab && cd ~/rebash-github-actions/docker-lab

Focus: Dockerfile plus build-push workflow (local build; push disabled)

Step 1 – Dockerfile and workflow

mkdir -p .github/workflows
cat > Dockerfile << 'EOF'
FROM python:3.12-alpine
WORKDIR /app
COPY app.py .
USER nobody
CMD ["python", "app.py"]
EOF
echo 'print("gha docker lab")' > app.py


```yaml
# .github/workflows/docker.yml
name: Docker pipelines
on: [push, workflow_dispatch]
permissions:
  contents: read
  packages: write
jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: docker/setup-buildx-action@v3
      - uses: docker/build-push-action@v6
        with:
          context: .
          push: false
          tags: ghcr.io/example/demo:lab
          # production tags often use: ghcr.io/${{ github.repository }}/demo:${{ github.sha }}

cat > .github/workflows/docker.yml << 'EOF' name: Docker pipelines on: [push, workflow_dispatch] permissions: contents: read jobs: build: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - uses: docker/setup-buildx-action@v3 - uses: docker/build-push-action@v6 with: context: . push: false tags: ghcr.io/example/demo:lab EOF 

### Step 2 – Local docker build proof

```bash
docker build -t rebash-gha-lab:local .
docker run --rm rebash-gha-lab:local
docker rmi rebash-gha-lab:local
grep -E 'build-push-action|push: false' .github/workflows/docker.yml

Final step – Cleanup note

docker rmi rebash-gha-lab:local 2>/dev/null || true
# Keep ~/rebash-github-actions/ for later tutorials

Validation

  • Lab commands run under ~/rebash-github-actions/module-07/.github/workflows/
  • You can explain each Theory section in your own words
  • You used modern tooling where it applies to this topic
  • You can describe one production failure mode for this topic

Code Walkthrough

Production practice for Docker Pipelines with GitHub Actions always combines:

  1. Inspect before you change (status, plan, logs, dry-run)
  2. Prefer reversible, documented changes (Git, IaC, drop-ins, version pins)
  3. Capture evidence (command output, pipeline logs) for handovers
  4. Prefer current tools and APIs over legacy shortcuts
  5. Least privilege — escalate credentials only when required

Keep runbooks short enough to follow under pressure. Automate checks; keep humans for judgement.

Security Considerations

  • Treat credentials and tokens for github-actions as privileged — never commit them
  • Prefer short-lived auth (OIDC, roles, SSO) over long-lived keys
  • Validate blast radius before apply/deploy/delete operations
  • Restrict who can approve production changes
  • Collect audit logs; limit who can read sensitive traces

Common Mistakes

Pushing latest from every pull request.

Validate assumptions against the Theory section and official docs before changing production.

Privileged Docker-in-Docker on shared self-hosted runners without isolation.

Lab shortcuts (open security groups, admin roles, skip approvals) must not ship unchanged.

Changing production without a rollback path

Always know how to revert (previous artefact, prior release, state rollback, DNS failback).

Best Practices

  • Encode Docker Pipelines with GitHub Actions changes as code and review them in pull requests
  • Pin versions (images, modules, actions, provider plugins)
  • Separate environments with clear promotion gates
  • Alert on symptoms with runbooks attached
  • Destroy lab resources; tag everything with owner and expiry where possible

Troubleshooting

Symptom Likely cause Fix
Auth / permission denied Wrong identity, policy, or scope Check caller identity, roles, and least-privilege policies
Timeout / no route Network, DNS, security group, or endpoint Trace path, DNS, and allow-lists before retrying
Drift / unexpected plan Manual change or wrong state/workspace Reconcile desired vs actual; avoid click-ops on managed resources
Pipeline/job red Flaky step, cache, or missing secret Read failing step logs; bisect recent workflow/config changes
Cost spike Idle load balancer, NAT, oversized compute Inventory billable resources; stop/delete labs promptly

Summary

Docker Pipelines with GitHub Actions is essential for Cloud and DevOps engineers working with github-actions. Practise the lab until the inspection and change path is muscle memory, then continue the track.

Interview Questions

  1. Why keep push false until registry auth is ready?
  2. How should image tags incorporate git SHA?
  3. What permissions are needed to push to ghcr.io?
  4. How does Buildx help multi-platform builds?
  5. How do you avoid storing registry passwords in the Dockerfile?

Sample answer — question 2

Check Dockerfile context, Buildx setup, and whether push/tags match registry permissions.

Sample answer — question 4

Use OIDC or GITHUB_TOKEN/registry login actions; never bake credentials into layers.

References