New: Roadmaps ordered paths through our cheat sheets and flashcards, so you always know what to study next.
Explore themSee what's new on GitHubFrom a Linux terminal to a production deployment strategy.
A 16-step learning path. Follow it in order, or jump to what you need.
This path is for developers, sysadmins, or support engineers ready to own how software gets built, shipped, and kept running. Plan on about 8 to 10 weeks at 3 to 5 hours a week, starting from a bare Linux terminal and ending at a real deployment strategy. It stays at generalist depth: for more depth on Kubernetes internals, Terraform modules, or on-call reliability practice, see the Docker-to-Kubernetes, Infrastructure as Code with Terraform, and Site Reliability Engineer paths instead. By the end you can build a CI/CD pipeline that ships code automatically, containerize and orchestrate an app with Docker and Kubernetes, and provision cloud infrastructure as versioned Terraform code.
No prior experience needed. Start from zero.
Open a terminal today and you can navigate, search, and script your own machine instead of clicking through a GUI, the one skill every later step in this path quietly assumes you already have.
Once a terminal feels normal, the next wall is jargon: DNS, ports, and subnets are the vocabulary every Docker network and cloud VPC later in this path is built from.
With Linux and networking vocabulary in place, you can start tracking every change you make and pushing it somewhere safe, the exact habit the pipeline in step 5 triggers off of.
You can move around Linux from the keyboard, read a network diagram without panicking, and save your work with Git instead of losing it. Next up: turning that workflow into a pipeline that ships on its own.
Finish this section to unlock.
+100 XP
Before opening a single pipeline tool, it helps to see the whole map: how the Git habit from step 3 connects through automation and provisioning to the monitoring still ahead.
Now that the map makes sense, walk through what actually happens between a commit and a running app: build, test, and deploy stages that turn manual releases into a repeatable sequence.
Pick a concrete tool to run those stages: GitHub Actions wires the pipeline stages from step 5 directly into the same Git workflow you picked up in step 3.
You can wire a GitHub Actions pipeline that builds, tests, and deploys on every commit, and a few minutes reviewing chapter one's due flashcards keeps Linux and Git sharp while you build on them here. Next up: packaging what that pipeline ships.
Finish this section to unlock.
+100 XP
Packaging an app so it behaves the same on your laptop and inside the pipeline you just wired up solves the oldest complaint in software: it worked on my machine.
Running one container is easy; keeping hundreds of them healthy is where most self-taught learners stall, and this is where orchestration finally clicks instead of feeling like memorized YAML.
Take this if you're aiming at a company still running Jenkins in production: the pipeline stages from step 5 carry over, only the syntax and plugin ecosystem change.
You can containerize an app with Docker and get Kubernetes to keep it running without babysitting it by hand, the point where most self-taught learners stall out. Next up: provisioning the infrastructure it all runs on instead of clicking through a console.
Finish this section to unlock.
+100 XP
Containers need somewhere to run at real scale, so before picking a provider, get the shared vocabulary, IaaS versus PaaS versus serverless, that every cloud console dresses up differently.
Clicking through a console doesn't scale past what step 10 just described, so the next move is describing that same infrastructure in files you can version the way you versioned code back in step 3.
Terraform turns the declarative ideas from step 11 into a real workflow, plan then apply, that provisions the servers and networks your step 7 and step 8 containers actually run on.
Choose this if your target role also configures servers directly instead of only provisioning them: Ansible automates the day-two setup that Terraform in step 12 leaves undone.
Go deep here if you want one cloud provider nailed down now instead of later: AWS turns the generic services from step 10 into concrete compute, storage, and networking you can actually click through.
Deployed apps fail quietly without visibility, so before shipping under pressure, learn to watch what step 6's pipeline and step 8's cluster are actually doing: the metrics, logs, and traces underneath both.
Everything up to now, pipelines, containers, provisioned infrastructure, and monitoring, comes together in the one decision that decides whether a release is safe: how you actually roll it out to production.
You can provision cloud infrastructure as versioned code, watch it in production, and pick a deployment strategy that ships safely under pressure, the full loop from a blank terminal to a resilient release. That's the DevOps Engineer badge, earned.
Finish this section to unlock.
+100 XP
Finish every required step, at least 70% of them genuinely done (not skipped), to earn this badge and 500 XP.