Start reading
Study notes on Designing Data-Intensive Applications and Kafka: The Definitive Guide — 29 chapters, 346 sections, 607 rendered diagrams.
Two books, read closely and written up chapter by chapter. Each chapter follows the same shape: the argument in order, diagrams redrawn from the figures, technology deep dives, a production failure catalog, a decision cheat sheet, worked arithmetic, and a self-test.
Kleppmann & Petrov · 2nd edition
Designing Data-Intensive Applications
Chapter-by-chapter notes on Designing Data-Intensive Applications, 2nd edition — concepts, diagrams, technology deep dives, failure catalogs, and self-tests.
Shapira, Palino, Sivaram & Petty · 2nd edition
Kafka: The Definitive Guide
Chapter-by-chapter notes on Kafka: The Definitive Guide — producers, consumers, internals, exactly-once semantics, operations, and stream processing.
Row-oriented OLTP database
PostgreSQL
PostgreSQL is a row-oriented database built for OLTP: many small, concurrent transactions that read and write individual rows safely. Covers setup, MVCC, WAL, indexing, the planner, replication, and operations.
Column-oriented OLAP database
ClickHouse
ClickHouse is a column-oriented database built for OLAP — scanning and aggregating billions of rows fast. Covers MergeTree internals, querying, ingestion, scaling, and operations.
How each chapter is laid out
| Section | What's in it |
|---|---|
| Concepts | The chapter's argument, in order, with key sentences preserved verbatim where the exact wording matters |
| Diagrams | Redrawn from the book's figures, plus extra diagrams for mechanisms described only in prose |
| Technology deep dives | Per system: what problem it solves · why the alternative wasn't enough · how it works internally · deployment · monitoring · scaling · backup · what breaks in production |
| Production failure catalog | Symptom → underlying mechanism, filterable |
| Decision cheat sheet | The choices you'll actually have to make, with the criteria |
| Worked examples | The arithmetic behind the trade-offs |
| Self-test | 20–49 questions per chapter, ending in an open design question, with progress tracked |
| Terminology · Forward links | Chapter vocabulary, and where each idea gets developed next |
The books supply the concepts and the failure modes; the deployment/monitoring/backup material is the practitioner layer they deliberately leave to you.