DevOps Engineer roadmap
From first Linux command to a monitored, automated production system.
9 to 12 months at 8 to 10 hours a week8 stages39 topicsBeginner friendly
All roadmapsDevOps is the practice of shipping software quickly and safely: automating builds, deployments and infrastructure, and watching what runs. This path builds those skills in the order you will actually need them, and every stage ends with something you can show an employer.
Where you will end up
A portfolio of one application that you containerised, deployed through an automated pipeline onto cloud infrastructure you defined as code, and monitored with alerts. That is the evidence junior DevOps, platform and cloud support hiring managers look for.
Before you start
- Comfort using a computer and installing software
- Basic programming ideas help, but you do not need to be a developer
Roles it leads to
- Junior DevOps Engineer
- Cloud Support Engineer
- Site Reliability Engineer (after more experience)
- Platform Engineer (after more experience)
- 1
Linux, networking and Git
6 to 8 weeks0 of 4 done
Work confidently on a Linux server, understand how a request travels across a network, and keep every change in version control.
Prove it
Automate a server chore
Write a Bash script that backs up a directory, rotates old copies and logs what it did. Put it on GitHub with a README that explains how to run it.
- 2
Cloud and your first deployment
5 to 6 weeks0 of 5 done
Run a real application on a cloud server that is reachable by name over HTTPS and sensibly locked down.
Prove it
Host an app on your own domain
Deploy a small web app to a cloud VM with SSH key login only, a non-root user, a minimal firewall, a domain and a valid HTTPS certificate.
- 3
Containers
4 to 5 weeks0 of 4 done
Package an application and its dependencies so it runs the same everywhere.
Prove it
Containerise a multi-service app
Package an app with its database in Compose. Keep the image small, run as non-root, add a health check, and document how to start it with one command.
- 4
CI/CD pipelines
4 to 5 weeks0 of 5 done
Make every code change build, test and ship itself without manual steps.
Prove it
Push to deploy
Build a pipeline that tests, builds a container image, pushes it to a registry and deploys it, triggered only by a merge to the main branch.
- 5
Infrastructure as code
4 to 5 weeks0 of 4 done
Describe infrastructure in files, review it like code, and rebuild it from nothing.
Prove it
Rebuild it from an empty account
Define the stage 2 infrastructure in Terraform, apply it from scratch, deploy your app, then destroy it. Run the plan in your pipeline for every change.
- 6
Kubernetes and GitOps
6 to 8 weeks0 of 4 done
Run containers at scale with self-healing, rolling updates and Git as the source of truth.
Prove it
Ship to a cluster from Git
Deploy your containerised app to a Kubernetes cluster with Helm or Kustomize, driven by a GitOps tool, with a rolling update and a rollback you have tested.
- 7
Monitoring and reliability
4 to 5 weeks0 of 4 done
Know when something breaks before your users tell you, and learn from every incident.
Prove it
Dashboard, alert and runbook
Add metrics, a dashboard and one meaningful alert to your app, then write a short runbook and run a game day where you break it on purpose and recover.
- 8
Security and cost
3 to 4 weeks0 of 4 done
Build safety and cost control into the pipeline instead of bolting them on after launch.
Prove it
Secure and budget the pipeline
Add dependency, image and secret scanning as required checks, replace any stored cloud keys with short-lived credentials, and set a budget alert on the account.
Certifications that fit
Optional, and never a substitute for projects. Exam details change, so confirm on the official page before you book.
- AWS Certified Cloud PractitionerFoundationalA gentle first credential if you choose AWS as your cloud.
- HashiCorp Terraform AssociateAssociateMultiple choice, online proctored, one hour. The current version is listed on HashiCorp's site.
- Certified Kubernetes Administrator (CKA)ProfessionalA hands-on, two-hour, performance-based exam. Take it after you have run real clusters.
- AWS Certified DevOps Engineer - ProfessionalProfessionalA later goal once you have real experience building pipelines on AWS.
The honest part
Nobody finishes this by watching videos. The milestones are the real curriculum: build each one yourself, break it, fix it, and write down what you learned. Time estimates assume steady weekly study; your pace may be faster or slower, and that is fine. A certificate helps you get an interview. A working project you can explain gets you the job.
Checked against
Written by CloudIskole and reviewed against the official sources below. Last verified November 2026. Spotted something out of date? Tell us and we will fix it.
- Google SRE Book
- GitHub Actions documentation
- Docker documentation
- Terraform documentation
- Kubernetes documentation
- Prometheus documentation
- OpenTelemetry documentation
These roadmaps are our own work. For another community view of learning paths, see roadmap.sh, which inspired the idea of a visual guide.


