All roadmaps

DevOps Engineer

Make shipping boring - in the best possible way.

The complete DevOps path: Linux, networking, scripting, Git, containers, CI/CD, one cloud done properly, infrastructure as code, Kubernetes, observability and reliability practice - with free resources throughout.

Your progress0/63 topics · 0%

Progress is saved in this browser.

10 stages63 topics₹0 to follow

Filled nodes are the focus for this track; outlined ones can wait. Tap a topic to open it and mark it done, learning or skipped.

DevOps Engineer in 2026Related roadmapsDocker roadmapGit and GitHub roadmapBackend roadmapSystem Design roadmapProve it as you goReading is half of it. This roadmap isbacked by a free, scored DevOps Engineercertification with a public verificationlink and a DevOps & Cloud Engineer jobsimulation on real tickets - evidence arecruiter can check, not a claim. Marktopics done, learning or skipped as yougo; your progress carries into theclassic view.Get certified - freeJob simulationClassic viewKnow the job before the toolsThe DevOps Engineer job in 2026Which skills to learnLinux, cloud,KubernetesCI/CD and IaCAI in the pipelineWhich role to prepare forDevOps, SRE,platformFresher vsswitcherSectors, cities and payWhat companies expectThe hiring barWhat interviewstestProof they verifyWhat gets you hiredA 90-day planProjects thatget readCertification and proofBe at home on any server.LinuxShell & filesystemProcesses & servicesText processingBash scriptingPackage & user managementTrace why a request didn'tarrive.NetworkingTCP/IP & DNSHTTP(S) & TLSFirewalls, proxies & loadbalancersDiagnostic toolsVersion everything; revieweverything.Git & CollaborationGit fluencyGitOps mindsetPackage anything so it runsanywhere.ContainersDocker fundamentalsComposeRegistries & image hygieneEvery merge tested; everydeploy repeatable.CI/CDPipeline designSecrets managementDeployment strategiesArtifact versioningArtifact managementOne cloud, understood properly.Cloud (AWS)Core compute & storageIAMVPC & networkingManaged servicesServerlessEnvironments you can recreatefrom a repo.Infrastructure as CodeTerraform fundamentalsState & modulesConfiguration managementRun and debug workloads on theindustry's orchestrator.KubernetesCore objectsConfig & secretsDebugging workloadsHelm basicsGitOps & service mesh(awareness)Know it's broken before userstweet it.Observability & ReliabilityMetrics & dashboardsLogs & tracingAlerting & SLOsIncident practiceProve itFresher or experienced?Both tracks hold every topic. On the Fresher track thefoundations are filled and the later stages are outlined - dothem when a project needs them. On the Experienced track thefoundations are outlined and the later stages are filled. The2026 section at the top applies to both.Prefer a list?The classic view shows the same 9 stages as a sequentialchecklist with time estimates. Your marks are shared betweenthe two views.Open the classic view

Everything on this roadmap

Every stage and topic on the map, in reading order. Click any topic to open it.

4.Version everything; review everything.

5.Package anything so it runs anywhere.

9.Run and debug workloads on the industry's orchestrator.

The DevOps Engineer roadmap, topic by topic

Every topic on the map in reading order, with what to learn and the resources we recommend. Main topics are open; open any subtopic to read it. Select a topic name to open it in the topic panel, where you can track your progress.

  1. 1. Know the job before the tools

    The DevOps Engineer job in 2026

    2026

    DevOps engineers make software deliverable: the pipelines, the infrastructure, the containers, the clusters and the monitoring that turn a repository into a running service. In 2026 the job is shaped by two reports: the CNCF's 2025 survey found 82 percent of container users running Kubernetes in production, up from 66 percent in 2023, with two thirds of organisations hosting generative AI models using Kubernetes for inference; DORA's 2025 study of nearly 5,000 professionals found AI acts as an amplifier - it magnifies strong delivery practices and exposes weak ones. The four nodes answer which skills, which role, what companies expect and what gets you hired.

    Which skills to learn

    Linux, networking, one cloud and Kubernetes are the floor; CI/CD and infrastructure as code are the daily work; and AI in the pipeline - assistants writing the manifests and the scripts, agents watching the alerts - is the 2026 layer that DORA says only pays off on top of sound practices. The sub-nodes order the three.

    Linux, cloud, Kubernetes

    Linux as a working environment - the shell, processes, permissions, networking, systemd, logs - then networking proper, DNS, TLS, load balancing and firewalls, then one cloud learned to the level of its compute, storage, identity and networking primitives, and then containers and Kubernetes. The CNCF 2025 survey's 82 percent production figure for Kubernetes, and Indian hiring analyses that find it in most DevOps postings, make it unavoidable; the World Economic Forum's 2025 report ranks networks and cybersecurity the second-fastest-growing skill area to 2030. Learn Docker before Kubernetes and the cloud's basics before either.

    CI/CD and IaC

    The daily work is automation: a pipeline in GitHub Actions or GitLab that tests, builds, scans and deploys; infrastructure as code with Terraform or OpenTofu; configuration and secrets managed, not pasted; environments that can be rebuilt from the repository; and observability - metrics, logs, traces, alerts - so you know before the customer does. DORA's 2025 report describes seven capabilities, from a clear AI stance to healthy data and small batches, that make AI adoption pay; every one of them sits on this foundation. A team's delivery speed and stability are the measures you will be judged by.

    AI in the pipeline

    AI

    AI tools now write much of the configuration a DevOps engineer used to type - manifests, pipelines, Terraform, runbooks - and agents are starting to handle the first response to alerts. DORA's 2025 finding is the discipline: AI amplifies, so a team with flaky tests and no observability gets faster at breaking things. The Stack Overflow 2025 survey has 80 percent of developers using AI tools and 46 percent distrusting the output, which for infrastructure means reviewing every generated change against a plan, running it in a non-production environment first, and never letting an agent apply to production without a human step.

    Which role to prepare for

    DevOps engineer, site reliability engineer, platform engineer and cloud engineer overlap heavily and differ in what they own: delivery, reliability, the internal platform or the cloud estate. The sub-nodes compare them, lay out the fresher and switcher routes, and say where demand sits by sector and city.

    DevOps, SRE, platform

    A DevOps engineer owns delivery - pipelines, environments, deployments - for one or more teams. A site reliability engineer owns reliability: service levels, incident response, capacity, toil reduction, usually at larger companies. A platform engineer builds the internal platform other engineers deploy onto. A cloud engineer owns the cloud estate and its cost. The CNCF 2025 survey's near-universal Kubernetes adoption means all four share the cluster; choose by whether you like shipping, keeping things up, building tools or running the estate.

    Fresher vs switcher

    Freshers rarely start as DevOps engineers; the usual entry is a support, operations, QA or junior developer role with a visible automation project, or a junior cloud role after a certification. HackerRank's 2025 report found early-career hiring the hardest hit, so a public repository with a working pipeline and a cluster is the lever. Switchers from system administration, networking, QA or backend development bring the half the job needs and should fill the other half. On this map freshers get the foundations filled and the orchestration stages outlined; experienced learners get the reverse.

    Sectors, cities and pay

    Demand sits in SaaS and product companies, fintech, e-commerce, the global capability centres and the IT services firms that run infrastructure for the world; Indian hiring analyses through 2026 put cloud and DevOps engineers in the top five most demanded roles, and Nasscom's 2026 review reports an industry of 315 billion dollars with 59.5 lakh employees. Bengaluru, Hyderabad, Pune and Chennai lead; remote roles for foreign companies pay in dollars for on-call discipline and clear English. Pay rises with incidents you have run, not with tools you have listed.

    What companies expect

    Companies expect a DevOps engineer to take a service from repository to production reliably: a pipeline, infrastructure as code, a container on a cluster, secrets handled, monitoring and alerts, and an incident story told honestly. The sub-nodes cover the bar, the interview and what gets verified.

    The hiring bar

    The bar is a service you can ship and keep up: a pipeline that tests and deploys from a merge, infrastructure in code with a plan reviewed before apply, a container on a cluster with resource limits and health checks, secrets in a vault, dashboards and alerts that catch problems first, and a written post-mortem for the last incident. The Work Trend Index found 66 percent of leaders unwilling to hire without AI skills; DevOps teams expect the tools used for configuration and the engineer accountable for every apply.

    What interviews test

    Expect a practical task - debug a failing pipeline, write a Dockerfile, fix a Kubernetes deployment, author a Terraform module - often with an AI tool allowed; Linux and networking questions; a scenario about an outage and what you would do first; and a walk through a system you run. HackerRank's 2025 report says 66 percent of developers prefer real-world assessments and that they are now standard. The strongest answers are stories: an incident, its cause, what changed so it cannot recur.

    Proof they verify

    Reviewers inspect the repository and the running system: Terraform with a plan output, a pipeline with real runs, a cluster with a deployed service and a dashboard, a runbook. That beats any list of tools; OneRoadmap's free DevOps certification with its public verification link beats a course badge; the roadmap's final stage names the external cloud and Kubernetes credentials that employers recognise. A note on what the assistant generated and what you corrected is now part of the proof.

    What gets you hired

    A deployed pipeline and cluster a reviewer can inspect, infrastructure defined in code in a readable repository, a verified certification and a clear account of how AI tools were used. The sub-nodes give the ninety-day plan, the projects that get read and the proof that counts.

    A 90-day plan

    Month one: Linux, networking and one cloud, with a service running on a virtual machine behind TLS. Month two: Docker and a pipeline that builds, tests and deploys it; Terraform for the environment; secrets managed. Month three: Kubernetes with the service deployed, monitored and alerting, a runbook and a post-mortem for a fault you injected, the DevOps certification, and twenty targeted applications a week. Keep the log of what the assistant drafted and what you changed before apply.

    Projects that get read

    Projects that get read are running systems: a three-tier application deployed to Kubernetes with a pipeline, infrastructure as code, monitoring and a documented failure drill; a cost-tagged cloud environment rebuilt from code in one command; an internal tool that automates an on-call toil. Each has a repository, real pipeline history and a readme on the failure modes. The CNCF 2025 survey's 82 percent production Kubernetes figure is the market telling you what to build it on.

    Certification and proof

    Proof is verifiable: OneRoadmap's free DevOps certification with a public link, a GitHub profile with pipelines that have run, infrastructure code with plans, and a short account of your AI workflow. Octoverse 2025 counts more than 180 million developers on GitHub; the DevOps engineers who get replies are the ones whose systems a reviewer can see running.

  2. 2. Be at home on any server.

    Linux

    Be at home on any server.

    Time: about 3 weeks. Topics: Shell & filesystem · Processes & services · Text processing · Bash scripting · Package & user management.

    Shell & filesystem

    Optional on the Experienced track

    Navigation, globbing, permissions and ownership - half of production issues are permissions issues.

    Processes & services

    Optional on the Experienced track

    ps, top, signals and systemd units - finding, inspecting and managing what's actually running.

    Text processing

    Optional on the Experienced track

    grep, sed, awk, cut and pipes - slicing logs and configs at the speed of thought.

    Bash scripting

    Optional on the Experienced track

    Variables, conditionals, loops, exit codes and cron - automating the toil away is the job description.

    Package & user management

    Optional on the Experienced track

    apt/yum, sudoers and SSH keys - provisioning a fresh box into a working server.

  3. 3. Trace why a request didn't arrive.

    Networking

    Trace why a request didn't arrive.

    Time: about 2 weeks. Topics: TCP/IP & DNS · HTTP(S) & TLS · Firewalls, proxies & load balancers · Diagnostic tools.

    TCP/IP & DNS

    Optional on the Experienced track

    IP addressing, ports, TCP vs UDP and DNS resolution - the layered model as a debugging checklist.

    HTTP(S) & TLS

    Optional on the Experienced track

    Requests, status codes, certificates and termination - what 'it's probably TLS' actually means.

    Firewalls, proxies & load balancers

    Optional on the Experienced track

    Security groups, reverse proxies (nginx) and L4/L7 balancing - the boxes between users and your app.

    Diagnostic tools

    Optional on the Experienced track

    curl, dig, ss, traceroute and tcpdump basics - evidence before hypotheses.

  4. 4. Version everything; review everything.

    Git & Collaboration

    Version everything; review everything.

    Time: about 1 week. Topics: Git fluency · GitOps mindset.

    Git fluency

    Branching, merging, rebasing and bisect - history as a debugging tool, not just a backup.

    GitOps mindset

    Config and infra in repos, changes via PR - the audit log and rollback story for free.

  5. 5. Package anything so it runs anywhere.

    Containers

    Package anything so it runs anywhere.

    Time: about 3 weeks. Topics: Docker fundamentals · Compose · Registries & image hygiene.

    Build: Containerise a full app. Dockerise an app + database with Compose, health checks and a one-command dev environment.

    Docker fundamentals

    Images vs containers, Dockerfiles and layer caching - small images, fast builds, no snowflakes.

    Compose

    Multi-service local stacks with volumes, networks and health checks that mirror production shape.

    Registries & image hygiene

    Tagging strategies, vulnerability scanning and multi-stage builds - images as governed artifacts.

  6. 6. Every merge tested; every deploy repeatable.

    CI/CD

    Every merge tested; every deploy repeatable.

    Time: about 2-3 weeks. Topics: Pipeline design · Secrets management · Deployment strategies · Artifact versioning · Artifact management.

    Pipeline design

    Build → test → scan → publish → deploy as code in the repo; fail fast, cache aggressively.

    Secrets management

    Never in git: secret stores, scoped tokens and rotation - the difference between an incident and a breach.

    Deployment strategies

    Rolling, blue-green and canary - and rollbacks you've actually rehearsed before you need them.

    Artifact versioning

    Immutable builds, semver tags and 'deploy what you tested' - reproducibility as policy.

    Artifact management

    Registries for images and packages (ECR, Artifactory, GitHub Packages) - provenance and retention as first-class concerns.

  7. 7. One cloud, understood properly.

    Cloud (AWS)

    One cloud, understood properly.

    Time: about 4 weeks. Topics: Core compute & storage · IAM · VPC & networking · Managed services · Serverless.

    Core compute & storage

    EC2, S3 and EBS - instances, object storage and the pricing model that shapes architectures.

    IAM

    Users, roles and least-privilege policies - the security model everything else in AWS hangs off.

    VPC & networking

    Subnets, route tables, security groups and NAT - private networks you can reason about.

    Managed services

    RDS, load balancers and autoscaling groups - letting the cloud carry the undifferentiated weight.

    Serverless

    Lambda-style functions and managed containers (ECS/Fargate) - when no-server beats your-server, and the cold-start fine print.

  8. 8. Environments you can recreate from a repo.

    Infrastructure as Code

    Environments you can recreate from a repo.

    Time: about 2-3 weeks. Topics: Terraform fundamentals · State & modules · Configuration management.

    Build: Infra from zero, by code. Terraform a small environment (network, compute, DB) and deploy your containerised app to it via CI.

    Terraform fundamentals

    Providers, resources, plan/apply and variables - declarative infrastructure with a diff you can review.

    State & modules

    Remote state, locking and reusable modules - collaborating on infra without stepping on each other.

    Configuration management

    Ansible at concept level: idempotent server configuration when you're not fully containerised.

  9. 9. Run and debug workloads on the industry's orchestrator.

    Kubernetes

    Run and debug workloads on the industry's orchestrator.

    Time: about 3-4 weeks. Topics: Core objects · Config & secrets · Debugging workloads · Helm basics · GitOps & service mesh (awareness).

    Core objects

    Optional on the Fresher track

    Pods, Deployments, Services and Ingress - the vocabulary of every k8s conversation.

    Config & secrets

    Optional on the Fresher track

    ConfigMaps, Secrets and probes - wiring apps to their environment the k8s way.

    Debugging workloads

    Optional on the Fresher track

    kubectl describe/logs/exec and reading events - the triage loop for CrashLoopBackOff and friends.

    Helm basics

    Optional on the Fresher track

    Charts and values - packaging k8s apps so deploys are parameterised, not copy-pasted.

    GitOps & service mesh (awareness)

    Optional on the Fresher track

    ArgoCD/Flux syncing clusters from git, and meshes (Istio/Linkerd) for traffic and mTLS - where platform teams go next.

  10. 10. Know it's broken before users tweet it.

    Observability & Reliability

    Know it's broken before users tweet it.

    Time: about 2-3 weeks. Topics: Metrics & dashboards · Logs & tracing · Alerting & SLOs · Incident practice · Prove it.

    Metrics & dashboards

    Optional on the Fresher track

    Prometheus and Grafana: the four golden signals and dashboards that answer questions.

    Logs & tracing

    Optional on the Fresher track

    Structured, centralised logs and distributed traces - following one request across services.

    Alerting & SLOs

    Optional on the Fresher track

    Alert on symptoms users feel, page rarely, and let error budgets arbitrate speed vs stability.

    Incident practice

    Optional on the Fresher track

    Runbooks, on-call hygiene and blameless postmortems - reliability as a habit, not a hero.

    Prove it

    Optional on the Fresher track

    Take the OneRoadmap DevOps certification to verify the fundamentals on your resume.

Prove it as you go

Reading is half of it. This roadmap is backed by a free, scored DevOps Engineer certification with a public verification link and a DevOps & Cloud Engineer job simulation on real tickets - evidence a recruiter can check, not a claim. Mark topics done, learning or skipped as you go; your progress carries into the classic view.

Fresher or experienced?

Both tracks hold every topic. On the Fresher track the foundations are filled and the later stages are outlined - do them when a project needs them. On the Experienced track the foundations are outlined and the later stages are filled. The 2026 section at the top applies to both.

Prefer a list?

The classic view shows the same 9 stages as a sequential checklist with time estimates. Your marks are shared between the two views.

Frequently asked questions

What does a DevOps engineer actually do?

A DevOps engineer makes software deliverable and keeps it running: the pipelines that test and deploy, the infrastructure defined in code, the containers and clusters, secrets, monitoring and alerts, and the response when something breaks at night. In 2026 the configuration is often drafted by an assistant; the engineer reviews every change against a plan and owns every apply.

DevOps, SRE, platform or cloud engineer?

They overlap and differ in what they own: DevOps owns delivery for teams, site reliability owns service levels and incidents, platform engineering builds the internal platform others deploy onto, cloud engineering owns the estate and its cost. All four share Kubernetes, which the CNCF 2025 survey found in production at 82 percent of container users. Choose by what you like doing.

Do I need to be a developer first?

You need to read and write code - scripts, pipelines, infrastructure as code - but not to have shipped products. Many DevOps engineers came from system administration, networking, QA or support; many others from development. Both halves are learnable; the job needs the combination.

Which cloud should I learn?

One, properly, to the level of its compute, storage, identity and networking primitives. AWS has the most jobs, Azure dominates enterprise and Microsoft shops, Google Cloud is strong in data and AI. The concepts transfer; a second cloud is weeks once the first is understood. Learn Docker before Kubernetes and the cloud's basics before either.

Is Kubernetes mandatory?

For most DevOps roles in 2026, yes. The CNCF's 2025 survey found 82 percent of container users running Kubernetes in production and two thirds of organisations hosting generative AI models using it for inference; Indian hiring analyses find it in most DevOps postings. Learn it after Docker, on a small cluster, with one real service deployed and monitored.

How long does it take to become job-ready?

Someone with Linux and a cloud already needs a few months to add pipelines, infrastructure as code, containers and a cluster with a deployed service; a beginner should expect longer and often enters through support or a junior cloud role. Nobody can promise a timeline. The evidence is a running system a reviewer can inspect and an incident you can narrate.

Will AI replace DevOps engineers?

AI tools write manifests, pipelines and Terraform, and agents are starting to handle the first response to alerts. DORA's 2025 report found AI amplifies: strong teams get stronger and weak practices get faster at breaking things. The engineer who reviews generated infrastructure against a plan, keeps humans in the apply step and runs the incident is more valuable, not less.

How do I prove DevOps skills without job experience?

A running system: a three-tier application deployed to Kubernetes by a pipeline, infrastructure defined in code with plan outputs in the repository, monitoring and alerts, a runbook and a post-mortem for a fault you injected. Add OneRoadmap's free DevOps certification with its public verification link and the recognised cloud or Kubernetes credential named in the roadmap's final stage.