Chapter Thirteen
Case Studies
Four workload shapes assembled from the earlier material — a web SaaS platform, a data and ML platform, a monolith migration, and a multi-tenant internal platform.
Concepts land harder when you see them assembled. These four case studies take the objects and patterns from the rest of the course and build concrete architectures from them.
Each one is a different workload shape — a customer-facing SaaS, a data and ML platform, a monolith being pulled apart, and an internal platform serving many teams — and each closes with the alternatives considered and what a second attempt would change.
Topics in This Chapter
Topic 62
Web SaaS Platform
A customer-facing multi-service app — ingress, stateless tiers, a managed database, autoscaling, and the release process behind it.
Topic 63
Data and ML Platform
Batch and GPU workloads, large stateful stores, and pipeline orchestration — where Kubernetes fits and where it strains.
Topic 64
Migrating a Monolith
Strangling a monolith onto Kubernetes incrementally — the order of operations, the traps, and how to keep shipping during it.
Topic 65
Multi-tenant Internal Platform
A platform team serving many product teams — tenancy boundaries, golden paths, guardrails, and the self-service line.