Delivery loop
Plan → code → build → test → release → deploy → operate → observe. Short feedback loops help teams detect problems earlier.
A guided starting point for DevOps, Azure, Terraform, and Git—with hands-on exercises, reliable references, and interview questions that test understanding, not memorization.
Use this as a roadmap, not a checklist. Build small projects and explain your decisions.
A sensible order
Learn the foundations first, then combine the tools into one end-to-end delivery workflow.
Understand delivery, shell, networking, containers, and operational thinking.
02 · CollaborationVersion changes, review code, and automate build and release workflows.
03 · CloudLearn identity, networking, compute, storage, governance, and reliability.
04 · AutomationModel infrastructure, review plans, manage state, and build reusable modules.
01 · Foundations
DevOps is a way of improving how teams build, deliver, and operate software. It combines culture, automation, feedback, and shared responsibility—not just a collection of tools.
Open the interactive playgroundPlan → code → build → test → release → deploy → operate → observe. Short feedback loops help teams detect problems earlier.
Continuous integration frequently merges and verifies changes. Continuous delivery keeps releases deployable; continuous deployment automatically releases qualifying changes.
Metrics show numeric trends, logs record events, and traces connect work across services. Use them together to investigate system behavior.
Build security into design and delivery: least privilege, secret management, dependency and image scanning, policy checks, and actionable remediation.
Take a small web app, store its code in Git, build and test it in CI, package it as a container, deploy it to a sandbox, and add a health check plus a basic dashboard. Document the rollback and cleanup steps.
02 · Cloud platform
Learn the shared cloud concepts before memorizing service names. Understand how identity, network boundaries, resource organization, reliability, and cost fit together.
Microsoft Learn: Azure trainingMicrosoft Entra ID provides identity capabilities. Azure RBAC authorizes actions on Azure resources at scopes such as management group, subscription, resource group, or resource. Prefer managed identities for supported workloads and grant least privilege.
Management groups organize subscriptions; resource groups collect related resources; tags help classify ownership and cost. Azure Policy evaluates or enforces organizational rules. Understand scope and inheritance.
Study virtual networks, address spaces, subnets, NSGs, routing, private endpoints, DNS, load balancing, and connectivity. Be able to trace a request from client to service and identify where traffic is allowed or denied.
Compare virtual machines, app hosting, containers, and managed Kubernetes based on control, operations, scale, and fit. Learn storage redundancy options and how access, encryption, backup, and recovery requirements influence design.
Learn regions and availability zones, backup and disaster recovery objectives, Azure Monitor and alerts, resource health, budgets, cost analysis, and right-sizing. Design for failure and define how recovery will be tested.
Draw a simple web application architecture with a network boundary, application tier, data tier, identity approach, monitoring, backup, and a cost owner. Explain which services are public, how administrators connect, and what happens if one zone fails.
03 · Infrastructure as code
Terraform describes infrastructure declaratively. The workflow compares configuration with current state and provider APIs to propose changes before applying them.
Official Terraform tutorialsterraform settings and providers configure behavior; resources manage infrastructure; data sources read existing information; variables parameterize; outputs expose useful results.
Format → initialize → validate → plan → review → apply. A plan is a proposed change set, not a guarantee that the remote operation will succeed unchanged.
State maps configuration to real objects. Use an appropriate remote backend, locking where supported, access controls, backups, and careful handling of sensitive values. Never commit secrets or state containing secrets.
Modules package reusable configuration with clear inputs and outputs. Understand dependencies, replacement behavior, lifecycle settings, version constraints, and the impact of refactoring addresses.
Create a small, low-cost sandbox resource with variables and outputs. Run formatting, validation, and a plan. Review every proposed action, apply only in a safe account, then destroy and verify cleanup. Next, extract repeated configuration into a module.
04 · Version control
Git records snapshots of a project. Understanding the working tree, staging area, commits, and branches makes collaboration and recovery much easier.
Pro Git bookWorking tree: current files. Index/staging area: selected changes for the next commit. Repository: committed history. git add stages; git commit records a snapshot.
A branch is a movable reference to a commit. Merge combines histories; a fast-forward moves the branch pointer when possible. Conflicts require you to choose and validate the intended result.
git revert creates a new commit that reverses changes. git reset moves a branch pointer and can rewrite local history. Avoid destructive commands on shared work unless you understand the impact.
Fetch downloads remote history; pull fetches and integrates; push publishes commits. Use focused branches, small commits, reviewable pull requests, and clear messages.
$ git status
Inspect changed files
$ git switch -c feature/my-change
Create a focused branch
$ git add path/to/file
Stage intentional changes
$ git diff --cached
Review exactly what is staged
$ git commit -m "Explain the change"
Record a useful snapshot
Interview preparation
Use a clear pattern: clarify requirements → describe the design → explain trade-offs → cover security and failure modes → say how you would verify it. For experience questions, describe context, your actions, and measurable results without sharing confidential details.
In continuous delivery, changes are kept ready for release, but a person or business process may approve production deployment. In continuous deployment, qualifying changes are released automatically after required checks pass.
Establish impact and recent changes; inspect pipeline stage, deployment events, application health, logs, metrics, and dependencies. Stabilize or roll back using the team’s procedure, communicate status, then identify root cause and preventive checks.
Configuration expresses desired infrastructure. State records Terraform’s mapping between configured resource addresses and real objects, plus relevant attributes. State is operationally sensitive and should be protected and shared through an appropriate backend.
Inspect conflicted files and the surrounding intent, edit to the desired combined result, remove conflict markers, run relevant tests, stage the resolution, and complete the merge or rebase. Ask the change owners when intent is unclear.
Start with requirements and threat boundaries. Apply least-privilege identity, managed identities where supported, network restrictions, secret protection, encryption, policy and logging controls, patching, backups, and tested recovery. Verify controls rather than assuming defaults.
Establish ownership and usage baselines, find idle or oversized resources, validate utilization and service objectives, test right-sizing, automate schedules where suitable, and monitor cost and reliability after each change. Treat savings estimates as hypotheses to verify.
A VM runs a guest operating system on virtualized hardware. Containers isolate processes while sharing the host kernel. Containers are often lighter to start, but the right choice depends on isolation, compatibility, operations, and workload requirements.
Use a concise STAR-style story: situation and impact, your responsibility, actions taken to restore service, how you communicated, and the follow-up that reduced recurrence. Be clear about what you personally did and what the team did.
Publish a small, sanitized project with a README, architecture diagram, Terraform plan examples, CI workflow, security considerations, monitoring approach, cost notes, and teardown steps. Explain trade-offs and lessons learned; never publish credentials, customer data, or proprietary configuration.
Trustworthy references
Start with official documentation for product details; use community roadmaps as a guide, then practice in a safe lab.
External resources are maintained by their respective publishers. Service names, interfaces, certification objectives, and pricing can change; confirm details in current official documentation.
Continue with safe, simulated command practice or head back to the DevOps 10X community.