Career Roadmap
Site Reliability Engineer: Zero to Hero
This roadmap teaches the engineering skills required to become a modern Site Reliability Engineer. You'll build deep Linux and systems knowledge, automate infrastructure with Python and Terraform, operate Kubernetes in production, design observability using Prometheus, OpenTelemetry, and eBPF, and learn the incident management practices used by high-performing engineering teams. Rather than focusing on certifications alone, this roadmap emphasizes the practical skills, projects, and operational judgment expected from professional SREs. Use ExamOS practice quizzes throughout the journey to measure your progress.
Who is this roadmap for?
This roadmap is ideal for DevOps Engineers, Systems Administrators, Cloud Engineers, and Software Engineers with production experience who want to transition into Site Reliability Engineering. It also provides a structured learning path for Platform Engineers building internal developer platforms and Cloud Architects responsible for highly available distributed systems.
Skills You'll Develop
Diagnose production issues using Linux internals, performance tools, and kernel-level tracing.
Deploy, operate, secure, and troubleshoot production Kubernetes clusters.
Build monitoring, logging, and distributed tracing with Prometheus, Grafana, OpenTelemetry, and modern observability tooling.
Build reliable automation with Python, Terraform, Git, and CI/CD pipelines.
Design SLOs, manage error budgets, lead incidents, communicate with stakeholders, and drive continuous reliability improvements.
Target Roles in this Roadmap
- Site Reliability Engineer: Designs, operates, and continuously improves reliable distributed systems.
- Senior Site Reliability Engineer: Leads production incident response, defines reliability standards, and mentors engineering teams.
- Platform Engineer: Builds internal developer platforms, deployment pipelines, and infrastructure automation.
- Observability Engineer: Designs telemetry platforms using Prometheus, OpenTelemetry, and modern monitoring stacks.
- Cloud Reliability Engineer: Specializes in highly available, multi-region cloud infrastructure and disaster recovery.
Typical Employer Categories
- Technology & SaaS Companies: Large-scale web platforms, developer tooling, cloud-native applications, and AI infrastructure.
- Financial Services: Banking, payment platforms, trading systems, and fraud detection services requiring continuous availability.
- E-Commerce & Logistics: High-volume retail platforms, supply chain systems, and global delivery networks.
- Healthcare & Life Sciences: Mission-critical clinical systems, patient platforms, and healthcare infrastructure.
- Cloud Providers & Managed Services: Public cloud platforms, hosting providers, and managed infrastructure services.
- Consulting & Professional Services: Organizations modernizing enterprise infrastructure, cloud platforms, and reliability practices.
The Certification Path
Recommended Path
| Cert | When | Why |
|---|---|---|
| Linux Foundation LFCS | Month 2-3 | Hands-on systems foundation. Optional but recommended for candidates without deep Linux experience. |
| CNCF CKA | Month 6-10 | Kubernetes operations depth. The strongest hands-on certification for production Kubernetes operations. |
| Prometheus Certified Associate | Month 8-10 | Observability expertise. Validates practical monitoring and observability knowledge. |
| HashiCorp Terraform Associate | Month 10-12 | Infrastructure automation. Demonstrates infrastructure automation skills used across every cloud. |
| AWS DOP-C02 or Google Cloud Professional DevOps Engineer | Month 12-16 | Platform-specific SRE credential. The certification most explicitly aligned with SRE principles. |
Specialization Path
| Cert | When | Why |
|---|---|---|
| GitHub Foundations (GH-900) | Month 1 | Optional. Helps validate Git workflows, but practical Git experience is more valuable than the certification. |
| AWS SAA-C03 or AZ-104 or ACE | Month 4-6 | Cloud platform foundation. Required before platform-specific DevOps certification. |
| CKS (Certified Kubernetes Security Specialist) | Month 14-18 | Kubernetes security depth. Adds security specialization to SRE skills. |
| AWS Solutions Architect Professional | Month 16-20 | Advanced architecture depth. For SREs moving toward architecture-adjacent senior roles. |
Milestones: Junior → Mid → Senior
| Level | Milestone | When |
|---|---|---|
| Entry Level | Linux proficiency, Git workflow, Bash and Python automation | Month 3 |
| Practitioner Level**" | CKA, observability stack, Terraform, and CI/CD automation | Month 10-12 |
| Specialist Level | Platform DevOps certification, incident leadership, reliability portfolio, and production architecture experience | Month 16+ |
Step 0 - Understanding Site Reliability Engineering
Understand what Site Reliability Engineering actually is before learning the tooling. SRE is an engineering discipline focused on building reliable systems through automation, measurable service objectives, and continuous improvement rather than reactive operations.
~1 month~1 month
Step 0 - Understanding Site Reliability Engineering
Understand what Site Reliability Engineering actually is before learning the tooling. SRE is an engineering discipline focused on building reliable systems through automation, measurable service objectives, and continuous improvement rather than reactive operations.
- Why Google created SRE and how the discipline evolved
- SRE vs DevOps: culture versus implementation
- Service Level Indicators (SLIs), Service Level Objectives (SLOs), Service Level Agreements (SLAs), and error budgets
- Reducing operational toil through automation
- Blameless postmortems and continuous improvement
- Reliability as a business decision rather than a technical target
- Communicating technical risk to engineering, product, and business stakeholders
Certifications
💡 Site Reliability Engineering is rarely an entry-level role. Most engineers transition from software engineering, systems administration, platform engineering, or DevOps after gaining production experience.
💡 Reliability is never free. Every increase in availability carries operational and financial cost, which is why error budgets exist.
💡 Learn the difference between SLIs, SLOs, and SLAs early. They appear in interviews far more often than specific tools.
🛠Project Ideas
- â–¸Analyze a public incident report from Cloudflare, GitHub, AWS, or Google Cloud. Identify the SLI, SLO, likely error budget impact, root cause, and remediation actions.
- â–¸Read Google's SRE book and write a one-page reliability policy for a fictional e-commerce checkout platform.
Step 1 - Linux, Git, and systems foundations
Build deep Linux systems knowledge because nearly every production platform ultimately runs on Linux. Containers, Kubernetes, cloud virtual machines, and observability tooling all depend on understanding the operating system underneath.
~2-3 months~2-3 months
Step 1 - Linux, Git, and systems foundations
Build deep Linux systems knowledge because nearly every production platform ultimately runs on Linux. Containers, Kubernetes, cloud virtual machines, and observability tooling all depend on understanding the operating system underneath.
- Linux architecture: boot process, systemd, processes, scheduling, memory management, swap, and the OOM Killer
- File systems: ext4, XFS, permissions, ownership, ACLs, mounts, and storage fundamentals
- Linux networking fundamentals: sockets, DNS resolution, TCP/IP, HTTP, SSH, and network troubleshooting
- Command-line tools: ps, top, htop, vmstat, iostat, ss, lsof, journalctl, systemctl, grep, awk, sed, and find
- Performance troubleshooting using strace, perf, and native Linux diagnostic tools
- Git fundamentals: repositories, branching strategies (GitHub Flow, trunk-based development, GitFlow), pull requests, and code reviews
- Bash scripting for operational automation
Certifications
💡 Linux is the foundation of modern infrastructure. Kubernetes nodes, cloud virtual machines, and most production services ultimately rely on Linux internals.
💡 Network+ is optional. Review TCP/IP, DNS, HTTP, and routing if those concepts are unfamiliar, but invest most of your time in Linux diagnostics.
💡 Git is a daily operational tool. Infrastructure should be version controlled, peer reviewed, and deployed through pull requests rather than manual changes.
💡 Learn to investigate problems from the operating system first. Restarting services without understanding the root cause creates unreliable systems.
🛠Project Ideas
- â–¸Build a Linux virtual machine, deliberately create CPU, memory, disk, and service failures, then diagnose each issue using standard Linux tools before fixing it.
- â–¸Write a Git pre-commit hook that validates Terraform, YAML, or Kubernetes manifests before allowing commits.
- â–¸Write a Bash script that monitors disk usage, system load, and failed services, then logs alerts when thresholds are exceeded.
Step 2 - Programming, automation, and cloud engineering
Build the automation skills that distinguish Site Reliability Engineers from traditional operations roles. Learn to write reliable Python code, automate infrastructure changes, understand cloud platforms, and build repeatable deployment pipelines. Every manual operational task should become a candidate for automation.
~3-4 months~3-4 months
Step 2 - Programming, automation, and cloud engineering
Build the automation skills that distinguish Site Reliability Engineers from traditional operations roles. Learn to write reliable Python code, automate infrastructure changes, understand cloud platforms, and build repeatable deployment pipelines. Every manual operational task should become a candidate for automation.
- Python for SRE: modules, virtual environments, logging, exception handling, argparse, requests, boto3/Azure SDKs, retry and exponential backoff patterns
- Working with APIs: REST fundamentals, authentication, pagination, rate limiting, JSON and YAML processing
- GitHub Actions, Jenkins, or Azure DevOps: pipelines, runners, artifacts, environments, approval gates, secrets management, and reusable workflows
- Cloud platform foundations: AWS (SAA-C03), Azure (AZ-104), or GCP (ACE). Focus on IAM, networking, compute, storage, monitoring, and managed databases.
- Infrastructure as Code fundamentals: Terraform providers, resources, modules, state files, variables, outputs, and remote backends
- Operational automation: scheduled jobs, health checks, automated remediation, notifications, and infrastructure validation
Certifications
💡 The programming bar for SRE is reliability, not algorithm complexity. Write code that is observable, fault tolerant, and easy to maintain.
💡 Cloud providers differ, but IAM, networking, storage, compute, monitoring, and automation concepts transfer across AWS, Azure, and GCP.
💡 Terraform should become your default approach for infrastructure provisioning. Manual console changes should be the exception rather than the rule.
💡 Every automation script should produce structured logs, handle failures gracefully, and support retries where appropriate.
🛠Project Ideas
- â–¸Build a Python service that monitors cloud resources, detects unhealthy instances, posts alerts to Slack or Teams, and automatically retries failed API calls.
- â–¸Deploy a three-tier application using Terraform and automate deployments through GitHub Actions with pull request validation and approval gates.
- â–¸Build a reusable Terraform module for networking and deploy it into separate development and production environments.
Step 3 - Containers, Kubernetes, and production operations
Learn how modern production platforms actually run. Containers package applications, but Kubernetes operates them at scale. This step focuses equally on architecture, troubleshooting, and day-to-day operational excellence.
~3-4 months~3-4 months
Step 3 - Containers, Kubernetes, and production operations
Learn how modern production platforms actually run. Containers package applications, but Kubernetes operates them at scale. This step focuses equally on architecture, troubleshooting, and day-to-day operational excellence.
- Docker fundamentals: images, containers, Dockerfiles, multi-stage builds, container networking, volumes, image optimization, and security best practices
- Kubernetes architecture: control plane, worker nodes, reconciliation loop, scheduler, etcd, kubelet, and container runtime
- Core workloads: Pods, ReplicaSets, Deployments, StatefulSets, DaemonSets, Jobs, CronJobs, and Helm
- Networking: Services, Ingress, DNS, NetworkPolicies, service discovery, and load balancing
- Storage: Persistent Volumes, Persistent Volume Claims, StorageClasses, CSI drivers, and Stateful applications
- Security fundamentals: RBAC, service accounts, Secrets, ConfigMaps, admission controllers, and least-privilege access
- Production troubleshooting: kubectl logs, describe, events, exec, debug containers, resource limits, CrashLoopBackOff, ImagePullBackOff, scheduling failures, DNS failures, and networking issues
- Cluster lifecycle: upgrades, node maintenance, draining workloads, backup and restore of etcd, and high availability concepts
Certifications
💡 CKA is one of the few certifications that accurately reflects day-to-day operational work because it is entirely hands-on.
💡 Kubernetes troubleshooting is a pattern-recognition skill. The fastest engineers systematically eliminate networking, storage, scheduling, configuration, and application issues rather than guessing.
💡 Learn to read Kubernetes Events before reading application logs. Events often explain why workloads never started.
💡 Understanding the reconciliation loop is more valuable than memorizing kubectl commands.
🛠Project Ideas
- â–¸Deploy a production-style microservices application with Helm. Configure health probes, autoscaling, persistent storage, ingress, and resource limits.
- â–¸Intentionally introduce DNS failures, resource exhaustion, failed deployments, and broken readiness probes. Diagnose every issue using Kubernetes tooling without rebuilding the cluster.
- â–¸Build a highly available Kubernetes cluster, perform a rolling upgrade, back up etcd, simulate node failure, and restore normal operations.
Step 4 - Observability and production debugging
Learn how to understand production systems without guessing. Modern observability combines metrics, logs, traces, profiling, and kernel telemetry to explain why systems behave the way they do. This is where Site Reliability Engineers spend much of their time during incidents.
~2-3 months~2-3 months
Step 4 - Observability and production debugging
Learn how to understand production systems without guessing. Modern observability combines metrics, logs, traces, profiling, and kernel telemetry to explain why systems behave the way they do. This is where Site Reliability Engineers spend much of their time during incidents.
- Prometheus: pull architecture, exporters, service discovery, recording rules, alert rules, and PromQL (rates, histograms, quantiles, summaries)
- Grafana: building actionable dashboards using the RED and USE methodologies, variables, annotations, and dashboard design principles
- Structured logging: log aggregation with Loki or ELK, correlation IDs, log levels, and query techniques
- OpenTelemetry: metrics, logs, traces, context propagation, automatic instrumentation, manual instrumentation, and vendor-neutral telemetry pipelines
- Distributed tracing: Jaeger or Grafana Tempo, latency analysis, dependency graphs, and root cause investigation
- eBPF observability: bpftrace, Pixie, Cilium, and kernel-level tracing without modifying application code
- Alert engineering: symptom-based alerting, burn-rate alerts, alert routing, deduplication, escalation policies, and reducing alert fatigue
💡 Prometheus Certified Associate validates metrics expertise, but production SRE work increasingly combines Prometheus with OpenTelemetry and distributed tracing.
💡 Dashboards should answer operational questions, not display every available metric. Every graph should support a troubleshooting decision.
💡 The best alerts indicate user impact rather than infrastructure noise. Alert on symptoms before causes whenever possible.
💡 eBPF is becoming a standard observability capability, but learn traditional Linux diagnostics first so you understand what eBPF improves.
🛠Project Ideas
- â–¸Instrument a microservices application with OpenTelemetry. Export metrics to Prometheus, traces to Jaeger or Tempo, and logs to Loki. Correlate all three signals during a simulated outage.
- â–¸Create Grafana dashboards for the four golden signals (latency, traffic, errors, saturation). Design alerts based on error budget burn rate instead of simple CPU thresholds.
- â–¸Use bpftrace or Pixie to investigate high CPU utilization caused by a production workload. Compare kernel-level visibility with traditional Linux tools.
Step 5 - Reliability engineering, incident response, and platform automation
Move beyond operating infrastructure to engineering reliable systems. This step focuses on incident management, resilience engineering, automation at scale, and the operational practices that distinguish senior Site Reliability Engineers.
~2-3 months~2-3 months
Step 5 - Reliability engineering, incident response, and platform automation
Move beyond operating infrastructure to engineering reliable systems. This step focuses on incident management, resilience engineering, automation at scale, and the operational practices that distinguish senior Site Reliability Engineers.
- Reliability engineering: defining SLIs, SLOs, SLAs, error budgets, and measuring service health
- Incident management: severity classification, incident command, communication plans, stakeholder updates, escalation procedures, and recovery coordination
- Blameless postmortems: root cause analysis, contributing factors, corrective actions, and long-term reliability improvements
- Chaos engineering: failure injection, resilience validation, game days, disaster recovery testing, and fault tolerance
- Platform automation: self-healing systems, automated remediation, scheduled maintenance, runbooks, and operational workflows
- Infrastructure lifecycle: immutable infrastructure, blue-green deployments, canary releases, rollback strategies, and progressive delivery
- Operational security: secret management, OIDC federation, policy as code, least privilege, and infrastructure governance
Certifications
💡 Senior SREs spend less time provisioning infrastructure and more time designing systems that recover automatically when failures occur.
💡 Every recurring operational task should become a candidate for automation. The objective is to eliminate toil rather than manage it more efficiently.
💡 Strong incident communication is as valuable as technical troubleshooting. Clear updates build trust during outages.
💡 Chaos engineering should begin in lower environments and gradually expand as operational maturity increases.
🛠Project Ideas
- â–¸Build a self-healing platform that automatically detects unhealthy services, restarts failed workloads, notifies engineers, and records every remediation action.
- â–¸Conduct a game day by deliberately introducing failures into a Kubernetes application. Run the incident, document communications, restore service, and publish a complete blameless postmortem.
- â–¸Implement blue-green or canary deployments with automated rollback based on application health metrics.
Step 6 - Platform engineering and cloud specialization
Apply everything you've learned to a production cloud platform. At this stage you should think beyond individual services and focus on designing resilient, automated platforms that developers can safely deploy to. The platform-specific DevOps certifications validate these real-world engineering skills.
~2-3 months~2-3 months
Step 6 - Platform engineering and cloud specialization
Apply everything you've learned to a production cloud platform. At this stage you should think beyond individual services and focus on designing resilient, automated platforms that developers can safely deploy to. The platform-specific DevOps certifications validate these real-world engineering skills.
- AWS track (DOP-C02): CodePipeline, CodeBuild, CloudFormation, Systems Manager, CloudWatch, EventBridge, multi-account architectures, deployment strategies, and operational excellence
- Google Cloud track (Professional Cloud DevOps Engineer): Cloud Operations Suite, SLO implementation, Cloud Deploy, Cloud Build, GKE operations, incident management, and reliability engineering
- Azure track (AZ-400): Azure DevOps, GitHub Actions, Azure Monitor, deployment slots, infrastructure automation, and release management
- Platform engineering principles: reusable infrastructure, internal developer platforms, golden paths, self-service provisioning, and platform APIs
- Progressive delivery: blue-green deployments, canary releases, feature flags, automated rollbacks, and deployment verification
- Multi-region and disaster recovery architectures: high availability, backup strategies, recovery objectives (RPO/RTO), and business continuity planning
Certifications
💡 Platform engineering is becoming a natural evolution of Site Reliability Engineering. Modern SRE teams increasingly build platforms that prevent operational problems instead of repeatedly solving them.
💡 Choose the cloud platform that matches your target employers. The underlying reliability principles remain the same across AWS, Azure, and Google Cloud.
💡 Certifications validate experience but cannot replace it. Build at least one complete production platform before scheduling your exam.
🛠Project Ideas
- â–¸Build a production-grade CI/CD platform that deploys applications automatically using progressive delivery with health-based rollback.
- â–¸Design a multi-region application with automated failover, infrastructure as code, monitoring, alerting, and documented disaster recovery procedures.
- â–¸Build an internal developer platform that provisions infrastructure, deploys applications, and configures observability through a self-service workflow.
Step 7 - Senior SRE skills, portfolio, and career readiness
Technical expertise gets you into SRE. Engineering judgment, communication, and leadership are what make you a senior engineer. Consolidate your work into a portfolio that demonstrates reliability engineering at production scale.
~1 month (ongoing)~1 month (ongoing)
Step 7 - Senior SRE skills, portfolio, and career readiness
Technical expertise gets you into SRE. Engineering judgment, communication, and leadership are what make you a senior engineer. Consolidate your work into a portfolio that demonstrates reliability engineering at production scale.
- Portfolio development: document infrastructure projects with architecture diagrams, runbooks, dashboards, postmortems, and operational metrics
- System design for reliability: identifying failure modes, eliminating single points of failure, designing for graceful degradation, and capacity planning
- Incident command: severity assessment, communication cadence, stakeholder management, decision making under pressure, and coordinating recovery
- Executive communication: translating technical failures into business impact, risk assessments, and recovery timelines
- Mentoring and operational leadership: improving engineering standards, reducing operational toil across teams, and building a culture of reliability
- Career development: interview preparation, architecture discussions, GitHub portfolio, technical writing, conference talks, and continuous learning
Certifications
💡 Hiring managers evaluate engineering judgment as much as certifications. A portfolio demonstrating real operational experience often outweighs another credential.
💡 Clear communication during incidents is one of the defining characteristics of senior Site Reliability Engineers. Practice explaining technical problems to both engineers and executives.
💡 Continue building projects even after landing your first SRE role. Reliability engineering evolves rapidly, and continuous learning is part of the profession.
💡 CKS is the natural next certification if your responsibilities expand into Kubernetes security and platform hardening.
🛠Project Ideas
- â–¸Build a complete SRE portfolio containing a Kubernetes platform, full observability stack, Terraform infrastructure, CI/CD pipeline, documented chaos experiment, and a detailed blameless postmortem.
- â–¸Run a simulated production incident with multiple participants. Lead the incident response, coordinate communications, restore service, and produce executive and engineering postmortem reports.
- â–¸Design a highly available global application. Document architecture decisions, SLOs, disaster recovery plans, capacity assumptions, and operational runbooks.
Final Step - What this path actually builds
This roadmap develops the skills required to operate modern production systems with confidence. You begin by mastering Linux, automation, and cloud infrastructure before moving into Kubernetes operations, observability, reliability engineering, and platform automation. Along the way, you learn how to investigate production failures using metrics, logs, traces, and kernel-level telemetry instead of guesswork. The certifications validate important milestones, but they are not the destination. Senior Site Reliability Engineers are recognized by their engineering judgment: they automate repetitive work, design resilient systems, define meaningful SLOs, lead incident response, communicate clearly under pressure, and continuously improve reliability through postmortems and automation. Build real systems, practice failure, and measure your readiness with ExamOS. Book your exams only when your operational reasoning is as strong as your technical knowledge.
Final Step - What this path actually builds
This roadmap develops the skills required to operate modern production systems with confidence. You begin by mastering Linux, automation, and cloud infrastructure before moving into Kubernetes operations, observability, reliability engineering, and platform automation. Along the way, you learn how to investigate production failures using metrics, logs, traces, and kernel-level telemetry instead of guesswork. The certifications validate important milestones, but they are not the destination. Senior Site Reliability Engineers are recognized by their engineering judgment: they automate repetitive work, design resilient systems, define meaningful SLOs, lead incident response, communicate clearly under pressure, and continuously improve reliability through postmortems and automation. Build real systems, practice failure, and measure your readiness with ExamOS. Book your exams only when your operational reasoning is as strong as your technical knowledge.
Certifications
Final Thoughts
💡 Total: 12-18 months at approximately 2 hours/day (9-12 months at 3-4 hours/day)
💡 Linux & Systems: 2-3 months. Build a strong operating system foundation before moving to Kubernetes.
💡 Automation & Cloud: 3-4 months. Python, Terraform, CI/CD, and one major cloud platform.
💡 Containers & Kubernetes: 3-4 months. Production operations and CKA preparation.
💡 Observability & Reliability: 3-4 months. Prometheus, OpenTelemetry, incident response, and SLOs.
💡 Platform Engineering: 2-3 months. Cloud DevOps certification, disaster recovery, and production automation.
💡 Readiness: Consistent 80%+ scores on Legend mode across five sessions indicate strong exam readiness.
Honest Timeline
| Path | Minimum Study Pace | More Realistic Pace |
|---|---|---|
| Linux + Kubernetes (CKA) | ~5 months | ~7-8 months |
| CKA + Observability + Terraform | ~8 months | ~10-12 months |
| Complete SRE Roadmap (from DevOps or Systems background) | ~10 months | ~14-16 months |
| Complete Roadmap (from software engineering or non-operations background) | ~12 months | ~18-24 months |