Architecting at Scale

The companion repository for the book Architecting at Scale (Packt): the per-chapter code, extended notes, and all 73 Architect's Prompts that go with the book.

ShopFlow is the running example throughout the book: an online store for home and lifestyle goods that begins as a monolith serving 100 orders a day and ends as a globally distributed, observable, resilient, cost-governed, AI-augmented system that keeps changing safely. Every chapter solves the problem the previous chapter's solution created; this repository holds the artifacts that go with that journey.

Layout

One folder per chapter, ch1 through ch16:

Path	What it holds
chN/prompts.md	The Architect's Prompts from that chapter, verbatim, each with the When to use this line
chN/*.md	Extended notes, the detail that did not fit on the page
chN/code/	Runnable samples for that chapter

And at the root:

Path	What it holds
PROMPTS.md	Index of all 73 Architect's Prompts across the book

Chapters

Ch	Title	Prompts	Notes
1	The Scalability Mindset - When and Why to Scale	2
2	Core Principles of Scalable Architecture	3
3	Security-First and Compliance-First Architecture	4
4	Scaling the Global Delivery Layer - Edge, CDNs, and Beyond	4
5	Scaling the Modern Web Application – State, Performance, and Micro-Frontends	5	6
6	Architecting Scalable Services – Decomposition and API Design	5
7	Scaling Service Infrastructure – Resilience, Mesh, and Compute	5
8	Event-Driven Scaling – Decoupling with Messaging	5
9	Caching Strategies – Faster and Cheaper Scaling	5
10	Scaling Data and Databases – Storage, Queries, and Beyond	5
11	Observability – Seeing and Understanding Your System	5
12	Resilience and High Availability – Designing for Failure	5
13	Performance Tuning and Capacity Planning	5
14	Cost Optimization and Efficiency (FinOps)	5
15	AI-First Architecture: Pragmatism Over Hype	5
16	Continuous Experimentation and the Future-Proof System	5

Using the prompts

Every prompt is reproduced exactly as it appears in the book. They are deliberately demanding: each one asks for a decision with the reasoning attached, not a summary. Two things to do before you run one:

1. Replace every [bracketed placeholder] with your own data: a module inventory, a telemetry series, a cost breakdown. The prompts are built to reason over real numbers, and they produce generic advice when given generic input.
2. Check the When to use this line. The prompts are diagnostic instruments, and reaching for the wrong one is how you get a confident answer to a question you did not have.

The ShopFlow stack

The book commits to specific technology so the trade-offs stay concrete. Where a chapter names a tool, the samples here use the same one:

Concern	Choice
Relational store	MySQL 8 (sharded by a fixed 1,024-bucket logical key from Chapter 10)
Cache	Redis (Chapter 9)
Search	Elasticsearch, fed by CDC (Chapter 10)
Object storage	S3-compatible, URL-only in the relational row (Chapter 10)
Messaging	Managed RabbitMQ, with Kafka as the measured migration target (Chapter 8)
Services	TypeScript / Node.js, six bounded contexts (Chapter 6)
Front end	React micro-frontends behind a shell (Chapter 5)
Observability	OpenTelemetry, traces, metrics, structured logs (Chapter 11)
Load testing	k6 (Chapter 2)
Agent governance	Runtime policy enforcement at a single chokepoint (Chapters 14–15)

Status

The prompts are complete for all 16 chapters. Code samples and extended notes are being added chapter by chapter; Chapter 5 has the first set of extended notes.