Cloud¶

A cloud platform is just someone else's API for compute, network, and storage. The primitives repeat across AWS, Azure, and GCP; the interview-relevant skill is knowing what each layer actually guarantees and where it silently stops guaranteeing anything.

This section covers the layer below Kubernetes (how your code becomes a runnable artifact and how that artifact gets to production) rather than vendor consoles.

The Path From Laptop to Production¶

flowchart LR
    C[Code] --> B[Build: Docker image]
    B --> R[(Registry)]
    R --> D[Deploy: CI/CD pipeline]
    D --> S[Deployment strategy]
    S --> K[Kubernetes]
    I[Terraform: cluster, VPC, DB, IAM] -.provisions.-> K
    I -.provisions.-> R
    style I fill:#6a1b9a,color:#fff

Four questions map to four pages:

Question	Page
How does my code become a runnable, portable artifact?	Docker
What's the general idea behind "infrastructure defined as text," before any specific tool?	Infrastructure as Code
How does the infrastructure that artifact runs on get created, and stay reproducible?	Terraform
How does a commit turn into a running artifact, automatically and safely?	CI/CD
How does a new version reach users without an outage?	Deployment Strategies
How does the artifact actually run and stay healthy at scale?	Kubernetes

Mental model

Docker answers "what am I shipping." Terraform answers "what does it run on." CI/CD answers "how does it get there." Deployment strategy answers "how do users stop seeing the old version without seeing an outage." Interviewers who ask "walk me through your deploy pipeline" are really asking whether you can keep these four concerns separate — candidates who collapse them into "we use Kubernetes" get follow-up questions until the gaps show.

Pages in This Section¶

Page	Covers
Cloud Provider Comparison	NEW — AWS/GCP/Azure mental models; VPC & Auth as foundations; service mappings across providers
Docker	Images vs. containers, layer caching, multi-stage builds, networking, volumes
Infrastructure as Code	Declarative vs. imperative, idempotency, the reconciliation loop, tool landscape
Terraform	State, plan/apply, modules, drift, blast radius
CI/CD	Pipeline stages, artifact promotion, GitOps, security gates
Deployment Strategies	Rolling, blue-green, canary, and 12 more — with the failure each one buys you out of
IAM & Managed Services	Vendor-mapped IAM, managed DB, and event-bus comparison across AWS/GCP/Azure
FinOps	Tagging, showback/chargeback, commitment models, rightsizing, cost anomaly debugging
Release Engineering	Semantic/API versioning, lockfiles and dependency hell, build-layer caching, cron vs. event-triggered scheduling
Kubernetes	Request path, probes, kubectl diagnosis

Cloud Provider Comparison is the entry point: it explains the mental model (VPC, Auth, then everything else is a derivative), maps AWS services to GCP/Azure equivalents, and teaches when to use each cloud.

Next: Docker →