Multi-Cloud DevOps Masterclass — 100 Hours of Terraform, Kubernetes, GitOps and DevSecOps

• Images, layers, registries, multi-stage builds, distroless and image slimming
• Networking, volumes, compose, healthchecks, resource limits
• Security: non-root users, read-only filesystems, scanning, SBOMs, signing (cosign)
• Buildx, caching in CI, reproducible builds
• Lab: shrink a 1.2 GB image to 90 MB and prove it still works

• Architecture: control plane, kubelet, scheduler, etcd, CRI/CNI/CSI
• Workloads: Pod, Deployment, StatefulSet, DaemonSet, Job, CronJob
• Config: ConfigMap, Secret, env, projected volumes; probes and lifecycle hooks
• Networking: Service types, Ingress, Gateway API, DNS, NetworkPolicy
• Storage: PV/PVC, StorageClass, CSI drivers, stateful workload patterns
• Scheduling: requests/limits, QoS, affinity, taints, topology spread, PDBs
• Lab: debug six deliberately broken clusters

• Managed Kubernetes compared: EKS vs AKS vs GKE (and GKE Autopilot)
• Autoscaling: HPA, VPA, Cluster Autoscaler, Karpenter
• Helm: charts, values, dependencies, hooks; Kustomize overlays
• RBAC, service accounts, workload identity, admission control, Pod Security Standards
• Upgrades, drain, blue/green and canary strategies, disaster recovery
• Lab: zero-downtime cluster upgrade with a live traffic test

• HCL, providers, resources, data sources, variables, outputs, locals, functions
• State: remote backends, locking, drift, import, moved/removed blocks
• Modules: composition, versioning, registry, testing with terraform test and Terratest
• Workspaces vs directories; Terragrunt; multi-account/multi-subscription/multi-project patterns
• Cloud-specific modules for AWS, Azure and GCP — the same architecture three ways
• Policy as code: Sentinel/OPA, cost estimation with Infracost, CI validation
• Lab: Project 1 three-cloud landing zone

• Ansible: inventories, playbooks, roles, idempotency, Ansible Vault
• Packer for golden images; immutable vs mutable infrastructure
• Cloud-init, bootstrapping and drift prevention
• Lab: build and bake an AMI/image with Packer + Ansible

• Pipeline design: build once, promote everywhere; artefact registries; versioning
• GitHub Actions: workflows, matrix builds, reusable workflows, OIDC to cloud (no long-lived keys)
• Azure DevOps: pipelines, environments, approvals, service connections
• GitLab CI and Jenkins essentials for brownfield estates
• Testing in the pipeline: unit, integration, contract, smoke, load
• Release strategies: blue/green, canary, feature flags, database migrations
• Lab: multi-environment promotion with manual approval and automated rollback

• GitOps principles; ArgoCD and Flux compared; app-of-apps and ApplicationSets
• Progressive delivery with Argo Rollouts / Flagger and SLO-based automatic rollback
• Secrets in GitOps: Sealed Secrets, External Secrets Operator, Vault
• Internal developer platforms: Backstage, golden paths, self-service
• Lab: Project 2 GitOps platform

• Metrics with Prometheus: exporters, PromQL, recording and alerting rules, Alertmanager
• Logs with Loki/ELK; traces with OpenTelemetry and Tempo/Jaeger
• Grafana dashboards that people actually use; RED and USE methods
• SLIs, SLOs, error budgets, on-call, incident response and blameless postmortems
• Cost observability across three clouds; FinOps basics
• Lab: define SLOs for the platform and wire burn-rate alerts

• Threat modelling; supply-chain security (SLSA), dependency and container scanning
• SAST/DAST, secret scanning, IaC scanning (tfsec/Checkov), policy gates
• Secrets management with Vault and cloud-native KMS; short-lived credentials
• Compliance basics: CIS benchmarks, audit trails, evidence collection
• Capstone review, Terraform Associate / CKA prep pointers, DevOps interview drill
• Resume framing for DevOps and SRE roles

You need at least one to complete every lab. Terraform modules are written so the same architecture deploys on all three, and we show the differences side by side. Most students do the labs on one cloud and read along on the others.

It covers the CKA curriculum in Modules 3–4 and includes cluster-troubleshooting drills, but the CKA is a performance exam — plan an extra 20–30 hours of killer.sh style practice on top.

Very. Modules 3–5 and 7–9 are exactly what you need to own Airflow on Kubernetes, containerised Spark jobs and Terraform-managed data infrastructure. There is a data-engineer bundle with the Databricks or AWS course.

Only enough to maintain what exists. We spend most of the CI/CD time on GitHub Actions and Azure DevOps, with a short Jenkins segment for brownfield reality.