20 GitHub Projects to Practice DevOps Hands-On
20 verified GitHub repos to practise DevOps: Linux, Docker, Kubernetes, CI/CD, GitOps, Ansible, Terraform, monitoring and DevSecOps, with a task for each.
Good GitHub projects for DevOps practice give you exercises or a working system to operate: the Linux Upskill Challenge, Docker's voting app, CKAD exercises, AWS and Azure sample microservices, GitHub's starter workflows, Flux and Argo Rollouts demos, Ansible for DevOps, kube-prometheus, Grafana's observability demo, and deliberately vulnerable apps such as Juice Shop for DevSecOps practice.
This list covers 20 public repositories and gives one concrete practice task for each. Every repository was checked on the GitHub API on 7 November 2026: all 20 exist, none is archived, and each description below matches the project's own README. The list shares no repositories with our two earlier lists, GitHub repositories for DevOps learning (roadmaps and question banks) and GitHub repositories with hands-on DevOps projects (demo apps such as Online Boutique and podinfo).
Key takeaways
- Use these repositories as a practice gym. Clone, run, break and fix them, then move to building your own portfolio projects.
- Each entry has a practice task with a clear finish line, so you know when you are done.
- Most entries run on a laptop with Docker, local VMs or a kind cluster. The EKS blueprints and the techiescamp AWS projects need a cloud account, and Jenkins-Zero-To-Hero is written for EC2, so those can cost money.
- The four DevSecOps repositories are intentionally insecure. Run them only on an isolated local machine or cluster, never next to real workloads or credentials.
- Star counts change daily and say little about quality. Check the last commit date and the README before investing time.
- When you are ready to build from scratch, follow our 10 DevOps projects to build your portfolio.
The 20 repositories at a glance
| # | Repository | Area | Level | Runs locally? | Last push seen |
|---|---|---|---|---|---|
| 1 | livialima/linuxupskillchallenge | Linux administration | Beginner | Needs a Linux server or VM | Sep 2026 |
| 2 | docker/getting-started-app | Docker basics | Beginner | Yes | Jan 2026 |
| 3 | dockersamples/example-voting-app | Compose, Swarm, Kubernetes | Beginner | Yes | Oct 2024 |
| 4 | dgkanatsios/CKAD-exercises | Kubernetes exercises | Intermediate | Yes (kind) | Oct 2026 |
| 5 | kodekloudhub/certified-kubernetes-administrator-course | Kubernetes administration notes | Intermediate | Yes (kind or VMs) | Sep 2026 |
| 6 | aws-containers/retail-store-sample-app | Microservices on containers | Intermediate | Yes (Docker Compose, kind) | Nov 2026 |
| 7 | Azure-Samples/aks-store-demo | Microservices on Kubernetes | Intermediate | Yes (Compose or kind) | Nov 2026 |
| 8 | actions/starter-workflows | GitHub Actions templates | Beginner | Runs on GitHub | Aug 2026 |
| 9 | jenkins-docs/simple-java-maven-app | Jenkins pipeline | Beginner | Yes (Jenkins in Docker) | Nov 2026 |
| 10 | iam-veeramalla/Jenkins-Zero-To-Hero | Jenkins to Argo CD pipeline | Intermediate | Partly (written for EC2) | Mar 2025 |
| 11 | fluxcd/flux2-kustomize-helm-example | GitOps with Flux | Intermediate | Yes (kind) | Jul 2026 |
| 12 | argoproj/rollouts-demo | Canary and blue-green | Intermediate | Yes | Oct 2026 |
| 13 | geerlingguy/ansible-for-devops | Ansible playbooks | Beginner to intermediate | Yes (Vagrant VMs) | May 2025 |
| 14 | aws-ia/terraform-aws-eks-blueprints | Terraform for EKS | Advanced | No (AWS account) | Nov 2026 |
| 15 | prometheus-operator/kube-prometheus | Cluster monitoring | Intermediate | Yes (kind) | Nov 2026 |
| 16 | grafana/intro-to-mltp | Metrics, logs, traces, profiles | Intermediate | Yes (Docker Compose) | Nov 2026 |
| 17 | juice-shop/juice-shop | App security testing | Intermediate | Yes (isolated) | Nov 2026 |
| 18 | bridgecrewio/terragoat | IaC misconfiguration scanning | Intermediate | Yes (scan only) | Jul 2025 |
| 19 | madhuakula/kubernetes-goat | Kubernetes security | Advanced | Yes (isolated cluster) | Apr 2026 |
| 20 | techiescamp/devops-projects | AWS project collection | Intermediate | No (AWS account) | Nov 2026 |
"Last push seen" is the pushed_at date reported by the GitHub API on 7 November 2026. The voting app and Jenkins-Zero-To-Hero have had fewer recent changes, but both still run as described and are widely used for practice.
Linux and containers
1. livialima/linuxupskillchallenge
A free course in 22 lessons (Day 0 to Day 21) on running a remote Linux server from the command line. It starts with creating a server on AWS, Azure, Google Cloud, DigitalOcean or local hardware, then covers SSH keys, sudo, package management, Apache, text tools, ss and ufw, cron and systemd timers, permissions, logs and logrotate, and simple scripting.
Practice task: complete all lessons on one VM and keep a lab notebook in a Git repository with the commands you ran each day. Finish line: you can rebuild the server from your notes in under an hour.
2. docker/getting-started-app
The sample to-do application used in Docker's own getting started guide. It is small on purpose, so all your attention goes to Dockerfiles, images, volumes, bind mounts and multi-container setups.
Practice task: write a Dockerfile from scratch without copying the guide's version, then add a MySQL container with Compose. Finish line: data survives docker compose down and up.
3. dockersamples/example-voting-app
A distributed voting app spread across several containers: a Python front end for voting, Redis as a queue, a .NET worker, PostgreSQL for storage and a Node.js results page. The repository includes a Compose file, a Swarm stack file and a k8s-specifications folder of Kubernetes manifests, so the same app teaches all three.
docker compose up
# vote on http://localhost:8080, results on http://localhost:8081
kubectl create -f k8s-specifications/
# vote on NodePort 31000, results on 31001
Practice task: deploy it to a kind cluster, then convert the raw manifests into a Helm chart with configurable replica counts. Finish line: scaling the worker with helm upgrade works without editing YAML by hand.
Kubernetes
4. dgkanatsios/CKAD-exercises
Exercises grouped by the CKAD curriculum areas (core concepts, multi-container pods, pod design, configuration, observability, services and networking, state persistence), plus sections on Helm and custom resource definitions. Each exercise has a hidden solution, so you can test yourself first.
Practice task: work through one section a day on a kind cluster using only imperative kubectl commands and kubernetes.io, as you would in the exam. Finish line: you solve a full section without opening any solution.
5. kodekloudhub/certified-kubernetes-administrator-course
Public notes from KodeKloud's CKA course, organised into sections such as core concepts, scheduling, logging and monitoring, application lifecycle management, cluster maintenance, security, storage, networking, installing a cluster the kubeadm way, troubleshooting and mock exams. The notes explain etcd, the API server, the scheduler, kubelet and kube-proxy, taints and tolerations, node affinity and static pods.
Practice task: use the cluster maintenance section to upgrade a kubeadm cluster by one minor version and to back up and restore etcd. Finish line: the restored cluster shows the objects you created before the backup. Note that the hands-on labs themselves are part of the paid course; the repository holds the notes.
6. aws-containers/retail-store-sample-app
A sample retail store built for learning container platforms on AWS. It has UI, catalog, cart, orders and checkout services written in Java, Go and Node.js, pluggable data stores (MariaDB or MySQL, DynamoDB, Redis), Prometheus metrics and OpenTelemetry tracing in every component, Helm charts and a load generator. The README marks it as educational and not for production.
Practice task: run it locally with Docker Compose, then on kind with the published Helm charts, and scrape its metrics with Prometheus. Finish line: a Grafana panel shows request rates per service while the load generator runs.
7. Azure-Samples/aks-store-demo
A group of containerized microservices built for Azure Kubernetes Service demos: an order service in JavaScript, a product service in Rust, a makeline service in Go, two Vue.js web front ends, simulated customers and workers, RabbitMQ for the order queue and a DocumentDB instance. An optional AI service adds generated product text.
The README shows how to run it locally with Docker Compose, and the repository includes an all-in-one Kubernetes manifest (aks-store-all-in-one.yaml).
Practice task: apply the all-in-one manifest to a kind cluster (it does not need Azure to run), then write a NetworkPolicy so only the order service can reach RabbitMQ. Finish line: a test pod outside the allowed set cannot connect to the queue.
CI/CD
8. actions/starter-workflows
The workflow templates GitHub shows when you click "New workflow" in a repository's Actions tab, grouped into CI, deployments, automation, code scanning and Pages. The repository is not accepting contributions at present, but GitHub still maintains it and the templates are the quickest way to see standard patterns for each language.
Practice task: pick the CI template for your language, add caching and a test matrix across two runtime versions, and reduce permissions to read-only. Finish line: the matrix runs green and the workflow uses no write permissions it does not need. Our list of GitHub Actions workflows for automation gives fifteen patterns to try next.
9. jenkins-docs/simple-java-maven-app
The companion repository for the Jenkins tutorial "Build a Java app with Maven". It holds a small Java app with unit tests that produce a JUnit XML report, an example Jenkinsfile in the jenkins folder and a delivery script.
Practice task: run Jenkins in Docker, fork the repository and write your own declarative Jenkinsfile with build, test (publishing JUnit results) and deliver stages. Finish line: Jenkins shows a test trend graph after three builds.
10. iam-veeramalla/Jenkins-Zero-To-Hero
A step-by-step guide by Abhishek Veeramalla, popular with learners in India, that installs Jenkins on an EC2 instance, uses Docker containers as build agents and builds an end-to-end pipeline that deploys to Kubernetes with Argo CD in a GitOps style.
Practice task: follow it, but open only port 8080 in the security group rather than allowing all traffic (the README itself suggests this), and run Jenkins agents as Docker containers. Finish line: a commit to the app repository updates the image tag in the manifests repository and Argo CD deploys it.
GitOps and progressive delivery
11. fluxcd/flux2-kustomize-helm-example
The Flux project's reference layout for managing a staging and a production cluster from one repository, with apps, infrastructure and clusters folders. Flux installs a demo app from a HelmRepository using HelmRelease objects and upgrades it automatically within a semver range. The README asks for Kubernetes 1.33 or newer and suggests kind for local tests.
flux bootstrap github \
--components-extra=source-watcher \
--context=staging \
--owner=${GITHUB_USER} \
--repository=${GITHUB_REPO} \
--branch=main \
--personal \
--path=clusters/staging
Practice task: fork it, bootstrap two kind clusters as staging and production, and change a Helm value only for production. Finish line: flux get helmreleases -A shows different values per cluster from a single commit history.
12. argoproj/rollouts-demo
The demo app and examples for Argo Rollouts. Examples cover canary, blue-green, canary analysis with the Prometheus metric provider, experiments, preview stack testing and Istio traffic splitting. Images come in colours, plus bad- variants with a high error rate and slow- variants with high latency, so you can watch a rollout fail on purpose.
kustomize build examples/canary | kubectl apply -f -
kubectl argo rollouts get rollout canary-demo --watch
kubectl argo rollouts set image canary-demo "*=argoproj/rollouts-demo:bad-yellow"
Practice task: run the canary analysis example with Prometheus and deploy a bad- image. Finish line: the analysis fails and the rollout aborts without manual action. Read how canary releases reduce risk for the theory.
Infrastructure as code
13. geerlingguy/ansible-for-devops
Examples from Jeff Geerling's book Ansible for DevOps, arranged by chapter: a first playbook, multi-server orchestration with Vagrant, LAMP and Drupal, Node.js apps and roles, dynamic inventory, an ELK stack, rolling and zero-downtime deployments behind HAProxy, security hardening and more. Many examples boot local VMs with Vagrant and VirtualBox. The README notes that some playbooks favour teaching over strict best practice.
Practice task: run the deployments-rolling example, then refactor it into a role with Molecule tests. Finish line: a second playbook run reports changed=0.
14. aws-ia/terraform-aws-eks-blueprints
A collection of Terraform patterns that build complete Amazon EKS clusters with add-ons already installed. The README is explicit that these are reference patterns to copy and adapt, not a Terraform module to call directly.
Practice task: read one pattern end to end and draw the resources it creates, then deploy the simplest pattern in a sandbox AWS account and destroy it the same day. Finish line: terraform destroy completes and your billing console shows no leftover resources. EKS clusters and NAT gateways are billed by the hour, so set a budget alert first.
Monitoring and observability
15. prometheus-operator/kube-prometheus
Jsonnet-based manifests that install the Prometheus Operator, highly available Prometheus and Alertmanager, node-exporter, blackbox-exporter, the Prometheus Adapter, kube-state-metrics and Grafana, with default dashboards and alert rules. The README includes a compatibility table between its release branches and Kubernetes versions.
kubectl apply --server-side -f manifests/setup
kubectl wait --for condition=Established --all CustomResourceDefinition --namespace=monitoring
kubectl apply -f manifests/
Practice task: install it on kind, add a ServiceMonitor for an app you deployed earlier and write one PrometheusRule that fires when a pod restarts more than three times in ten minutes. Finish line: you see the alert in Alertmanager after crashing the pod. Our guide to Prometheus and Grafana together helps with the dashboards.
16. grafana/intro-to-mltp
Companion code for Grafana's introduction to metrics, logs, traces and profiles. One docker compose up starts Grafana, Mimir, Loki, Tempo, Pyroscope, k6, Beyla and Grafana Alloy with sample microservices, and an alternative Compose file uses the OpenTelemetry Collector instead of Alloy.
Practice task: start the stack, find a slow request in Tempo, jump to the matching logs in Loki and the profile in Pyroscope. Finish line: you can explain where the time in that request went.
DevSecOps (run in isolation)
The next three repositories are vulnerable by design so you can practise finding and fixing problems with defensive tools. Run them on a local machine or a throwaway cluster with no real credentials, and never expose them to the internet.
17. juice-shop/juice-shop
OWASP Juice Shop, an OWASP flagship project, is an intentionally insecure web shop that contains flaws from the whole OWASP Top Ten. It runs from a single container bound to localhost.
docker run --rm -p 127.0.0.1:3000:3000 bkimminich/juice-shop
Practice task: add a DAST stage to a CI pipeline that starts Juice Shop as a service container and runs an OWASP ZAP baseline scan against it, then publish the report as an artifact. Finish line: the pipeline produces a ZAP report on every run. Our OWASP ZAP in CI/CD guide walks through this.
18. bridgecrewio/terragoat
Terraform code for AWS, Azure and GCP that contains common cloud misconfigurations on purpose. Its README warns against deploying it in production or next to sensitive resources.
Practice task: do not apply it. Run a static scanner such as Checkov, Trivy (trivy config) or tfsec against the code, pick five findings and fix them in a branch. Finish line: the scanner report for your branch shows those five checks passing.
19. madhuakula/kubernetes-goat
An intentionally vulnerable Kubernetes environment with guided scenarios such as sensitive keys in codebases, NodePort services exposed by mistake and RBAC misconfiguration, plus defensive scenarios with Falco, Kyverno and CIS benchmark checks. The README asks you to run it only in a safe, isolated cluster.
Practice task: install it on a dedicated kind cluster, read each scenario's write-up and then apply the defensive fix: Pod Security Admission at the restricted level, NetworkPolicies and least-privilege RBAC. Finish line: after your changes, a kube-bench or Kyverno policy report shows fewer failed checks than before.
Project collections
20. techiescamp/devops-projects
A set of real-world style projects from the DevOpsCube team, each with automation scripts that provision infrastructure on AWS. Topics include a highly available Jenkins setup, Consul service discovery, a scalable Java app, a Prometheus stack with Docker Compose, a Terraform VPC design, client VPNs, Fargate deployments, GitHub Actions OIDC to AWS and Route 53 private hosted zones.
Practice task: do the GitHub Actions OIDC project, so your pipelines never store AWS keys. Finish line: a workflow assumes an IAM role and lists an S3 bucket with no secrets configured in the repository. The README also promotes certification discount codes; you can ignore that section.
Two other large collections, NotHarshhaa/DevOps-Projects and iam-veeramalla/aws-devops-zero-to-hero, were also active and not archived when checked. They overlap with the entries above, so pick one collection rather than all three.
An eight-week practice plan
| Week | Repositories | Goal for the week |
|---|---|---|
| 1 | 1 | Linux server basics, SSH, services, logs |
| 2 | 2, 3 | Dockerfiles, Compose, first Kubernetes manifests |
| 3 | 4, 5 | Kubernetes objects, scheduling, cluster maintenance |
| 4 | 8, 9 | GitHub Actions and Jenkins pipelines |
| 5 | 6 or 7, 15 | Run a microservices app and monitor it |
| 6 | 11, 12 | GitOps with Flux and canary releases |
| 7 | 13, 14 (read), 16 | Ansible roles, Terraform patterns, tracing |
| 8 | 17, 18, 19 | Security scanning and hardening in isolation |
How to practise without wasting time
- Fork before you change anything, so your work shows on your own profile with a commit history.
- Keep a
NOTES.mdin each fork with what you did, what broke and how you fixed it. These notes become interview answers. - Prefer local clusters (kind, minikube, k3d) and stop them when you finish. For cloud work, use a separate sandbox account with a budget alert.
- Read the repository's licence and README warnings before running it, especially for the security repositories.
- Once a practice task feels easy, rebuild the same thing in your own repository with your own app. That is when practice becomes a portfolio project.
Practising with a trainer
If you would like guided labs and someone to review your pipelines and clusters, our DevOps training in Pune uses hands-on exercises across Linux, Docker, CI/CD, Terraform, Ansible, Kubernetes and monitoring, in the Kharadi classroom or live online across India. For Kubernetes-heavy practice like entries 4 and 5, see CKA training. Learners who want to contribute upstream afterwards can use our list of DevOps open-source projects to contribute to.
Frequently asked questions
Which GitHub project should a DevOps beginner start with?
Start with the Linux Upskill Challenge for server basics, then Docker's getting-started app and the example voting app. Together they cover Linux, Dockerfiles, Compose and a first look at Kubernetes manifests without needing a cloud account.
How is this list different from your GitHub repositories for DevOps learning?
That list covers roadmaps, lessons and question banks for deciding what to study. A second list covers demo apps for portfolio projects. This one is a practice gym of exercises, labs and broken-on-purpose systems, and it shares no repositories with either.
Are these repositories free to use?
The repositories are free and open on GitHub under their own licences. Running them locally costs nothing. Deploying the AWS, Azure or EKS examples to a cloud account is billed by the provider, so use free tiers, set a budget alert and destroy resources after each session.
Is it safe to run Juice Shop, TerraGoat and Kubernetes Goat?
Yes, if you isolate them. Bind Juice Shop to localhost, scan TerraGoat without applying it, and run Kubernetes Goat on a throwaway local cluster with no real credentials. Never deploy them in a shared or production environment.
Can I put these practice repositories on my resume?
List them as practice, not as your own projects. What you can claim is the work you added in your fork, such as a Helm chart for the voting app or a ZAP stage for Juice Shop. Your own built projects carry more weight.
How do I check whether a GitHub repository is still maintained?
Look for an "archived" banner at the top of the page, the date of the latest commit, recent releases and whether issues and pull requests get replies. The GitHub API field archived and the pushed_at date give the same information in a script.
Do I need a powerful laptop for these labs?
Most run on 8 GB of RAM. kube-prometheus, the Grafana observability demo and the retail store sample are more comfortable with 16 GB. If your machine struggles, run one lab at a time and delete kind clusters you are not using.
Which of these help with Kubernetes certification?
CKAD-exercises maps directly to the CKAD curriculum areas, and the KodeKloud CKA notes follow the CKA syllabus. Practise on a kind cluster using kubernetes.io as your only reference, which mirrors exam conditions.
References
- livialima/linuxupskillchallenge
- docker/getting-started-app and Docker getting started guide
- dockersamples/example-voting-app
- dgkanatsios/CKAD-exercises
- kodekloudhub/certified-kubernetes-administrator-course
- aws-containers/retail-store-sample-app
- Azure-Samples/aks-store-demo
- actions/starter-workflows
- jenkins-docs/simple-java-maven-app and Jenkins tutorial: Build a Java app with Maven
- iam-veeramalla/Jenkins-Zero-To-Hero
- fluxcd/flux2-kustomize-helm-example
- argoproj/rollouts-demo
- geerlingguy/ansible-for-devops
- aws-ia/terraform-aws-eks-blueprints
- prometheus-operator/kube-prometheus
- grafana/intro-to-mltp
- juice-shop/juice-shop and OWASP Juice Shop project page
- bridgecrewio/terragoat
- madhuakula/kubernetes-goat
- techiescamp/devops-projects
- GitHub REST API: get a repository (archived, pushed_at)
- Kubernetes docs: Pod Security Admission
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0