Goal
An architect who cannot reason about production is designing fiction. This path builds the platform layer bottom-up — what a container is, what Kubernetes actually assembles, how infrastructure becomes reviewable text, how software ships continuously and stays observable. The house method applies double here: this domain is where cargo-culting is most rampant, and building each layer by hand is the antidote.
Prerequisites
- Foundations modules 4–5 (the OS and the network) — containers and clusters are those two subjects in a trench coat.
Modules
1. The cloud, provider-agnostic
- Read: Cloud fundamentals.
- Do: the note's drills — a small workload stood up with IAM, VPC, and managed data done properly; the same bill estimated before and read after.
- Proof: least-privilege IAM defended role by role; the cost model used as a design input, not a surprise.
2. Containers, from syscalls up
- Read: Kubernetes & containers deep dive (the container half) and Linux system deep dive as the reference under it.
- Do: Build your own container runtime.
- Proof: the subject's stage gates — namespaces, pivot_root, cgroup limits, veth
networking, an OCI image pulled and run;
docker runexplained flag by flag.
3. Kubernetes, assembled by hand
- Read: Kubernetes & containers deep dive (the orchestration half).
- Do: Kubernetes the Hard Way.
- Proof: the per-stage verification commands pass; when a pod fails to schedule you know which control-plane piece to interrogate first, and why.
4. Infrastructure as code
- Read: Infrastructure as code.
- Do: the note's drills — module 1's workload rebuilt declaratively, a drift introduced by hand and reconciled, state managed like the shared mutable resource it is.
- Proof: a colleague could recreate your environment from the repo alone; the plan/apply diff is your review surface.
5. Continuous delivery
- Read: CI/CD & delivery engineering.
- Do: the note's drills on a real project — trunk-based flow, a pipeline with quality gates, a feature flag driving a progressive rollout. The RealWorld capstone's milestone 5 is this module's natural workbench.
- Proof: tag-to-production with no human steps; a bad release rolled back (or flagged off) in minutes, demonstrated on purpose.
6. Observability & SRE practice
- Read: Observability & SRE practice.
- Do: the note's drills — instrument one of your builds with metrics, logs, and traces; define an SLO and burn its error budget in a game day.
- Proof: a mystery slowdown diagnosed from telemetry alone; the SLO document a teammate could operate from.
Related
- The path to excellence — the master path; Stage 4 alongside systems at scale.
- Security — TLS, secrets, and supply-chain hygiene thread through every module here.
- System Administration, Platform — the themes drawn on.