After 18 years in infrastructure and speaking at KubeCon and Red Hat Summit, here is the DevOps roadmap I would follow if starting today.
Phase 1: Foundations (Months 1-3)
Linux Fundamentals
You cannot do DevOps without Linux. Learn:
- File system navigation, permissions, processes
- Bash scripting, cron jobs, systemd
- Package management (apt, dnf)
- Networking: TCP/IP, DNS, firewalls, SSH
Recommended: Red Hat Enterprise Linux 9 for SysAdmins
Git and Version Control
- Branching strategies (trunk-based, GitFlow)
- Pull requests, code reviews
- Git hooks, signed commits
- Monorepo vs multi-repo
Networking
- HTTP/HTTPS, TLS certificates
- Load balancing, reverse proxies
- DNS resolution, CDNs
- VPN, Tailscale, WireGuard
Phase 2: Containers and CI/CD (Months 4-6)
Docker
- Building images (multi-stage builds)
- Docker Compose for local development
- Container security scanning
- Registry management (Harbor, GHCR, ECR)
CI/CD Pipelines
- GitHub Actions (most popular in 2026)
- GitLab CI, Jenkins (enterprise)
- Pipeline as code, reusable workflows
- Automated testing, security scanning
Infrastructure as Code
- Terraform for cloud provisioning
- Ansible for configuration management
- Both together — see Terraform vs Ansible
Phase 3: Kubernetes (Months 7-9)
Core Kubernetes
- Pods, Deployments, Services, Ingress
- ConfigMaps, Secrets, RBAC
- Helm charts, Kustomize
- Kubernetes Cheat Sheet
Production Kubernetes
- Multi-tenancy, resource quotas
- Network policies, service mesh (Istio, Cilium)
- Persistent storage (CSI drivers)
- Cluster autoscaling
GitOps
- ArgoCD or Flux
- Progressive delivery (canary, blue-green)
- Policy as code (Kyverno, OPA)
Phase 4: Cloud and Observability (Months 10-12)
Cloud Platforms
Master at least one:
- AWS (market leader)
- Azure (enterprise, strong K8s)
- GCP (best Kubernetes experience)
Observability Stack
- Metrics: Prometheus + Grafana
- Logs: Loki, OpenSearch, ELK
- Traces: Jaeger, Tempo, OpenTelemetry
- Alerting: PagerDuty, Opsgenie
Security (DevSecOps)
- Container image scanning (Trivy, Grype)
- SAST/DAST in CI pipelines
- Secret management (Vault, External Secrets)
- Supply chain security (Sigstore, SBOM)
Phase 5: AI and Platform Engineering (2026+)
AI Infrastructure
AI workloads are the fastest-growing Kubernetes use case:
- GPU scheduling (NVIDIA GPU Operator)
- Model serving (KServe, vLLM, llm-d)
- MLOps pipelines (Kubeflow, MLflow)
- AI on Kubernetes in Production
Platform Engineering
The evolution of DevOps:
- Internal Developer Platforms (IDPs)
- Self-service infrastructure
- Developer portals (Backstage)
- Golden paths and templates
Tools That Matter in 2026
| Category | Must-Know | Nice to Know |
|---|---|---|
| Containers | Docker, Kubernetes | Podman, containerd |
| IaC | Terraform, Ansible | Pulumi, Crossplane |
| CI/CD | GitHub Actions | GitLab CI, Dagger |
| GitOps | ArgoCD | Flux |
| Observability | Prometheus, Grafana | Datadog, Dynatrace |
| Security | Trivy, Vault | Falco, Kyverno |
| AI/ML | GPU Operator, KServe | Ray, Kubeflow |
| Networking | Cilium | Istio, Linkerd |
Certifications Worth Getting
- CKA (Certified Kubernetes Administrator) — the gold standard
- CKS (Certified Kubernetes Security Specialist) — differentiator
- AWS Solutions Architect or Azure Administrator
- RHCE (Red Hat Certified Engineer) — exam prep guide
- Terraform Associate
Career Progression
| Level | Focus | Years |
|---|---|---|
| Junior | Linux, scripting, CI/CD | 0-2 |
| Mid | Kubernetes, IaC, cloud | 2-4 |
| Senior | Architecture, security, mentoring | 4-7 |
| Staff | Strategy, platform design, org-wide | 7+ |
| Principal | Industry leadership, speaking, writing | 10+ |
My Advice
- Build things. Theory without practice is useless. Deploy a real cluster, break it, fix it.
- Contribute to open source. It is the fastest way to learn and build reputation.
- Write about what you learn. Blogging forces deeper understanding.
- Attend conferences. KubeCon changed my career trajectory.
- Do not chase every tool. Master fundamentals first.
Related Resources
- Ansible Playbook Examples
- Kubernetes Cheat Sheet
- Platform Engineering Technology Radar 2026
- My Books — 8 technical books on Ansible, Kubernetes, and RHEL
- Training and Workshops
