Architecture Weekly Issue #177 . Articles, books, and playlists on architecture and related topics. Split by sections, highlighted with complexity: 🤟 means hardcore, 👷‍♂️ is technically applicable right away, 🍼 - is an introduction to the topic or an overview. Now in telegram and Substack as well.
Highlights
Trade-off Analysis, Microservices Myths and the Origin of "The Hard Parts" 👷‍♂️
Last week my interview with Neal Ford went live. We discussed where the book came from, what is the hardest part in architecture and busted a couple of microservices myths. Grab yourself a hot drink and enjoy the conversation!
#interview #video
Shrink It to 14 kB — the TCP sweet-spot 🤟
The article argues that due to the nature of TCP slow start which is limited by 14 KB, there is a huge reason to have your webpage under this limit, as adding a single KB above that hits you with almost a half second. Find the technicalities inside.
#performance
A Crowd-Sourced Reliability Glossary 🤟
Antithesis have opened an A-to-Z of distributed-systems reliability terms—concise definitions of “Byzantine fault,” “gray failure,” “coordinated omission,” “write amplification,” and dozens more. Each entry cites a canonical paper and tags the relevant failure mode so post-mortem debates can skip the semantics and jump straight to root cause. Even better, the repo accepts pull-requests, letting ops teams contribute new jargon as it appears.
#distributedsystems
Wanna know how to Design Systems that deliver business value?
Business Oriented System Design Course Cohort #6 is officially open!
Looking for a way to advance your career? Felt you overgrew the mere feature development, but lack skills to design complete systems? Want to make the business impact? 10 hours of content packed lectures, engaging practice and the final work you will be proud to showcase as well as Credly(by Pearson)-based digital certificate proving your experience. More than 70 engineers already passed the course with amazing feedback and advanced their careers. New cohort starts on 23rd of July. Find the Details, Feedbacks and Enrollment into the course here. Only a single place left!
Follow-Up
LISTEN/NOTIFY: Postgres’ Surprise Bottleneck 👷‍♂️
Recall.ai learned the hard way that every NOTIFY grabs a global commit lock, stalling thousands of otherwise unrelated writers. Under 1 k messages / s their meeting-recording pipeline saw P95 latency explode from 3 ms to 1.8 s. A switch to Redis Streams and a Rust side-car restored headroom to 35 k msg/s and single-digit-ms writes. Follow the technical investigation inside.
Written By:Elliot Levin
#db #performance #postgres
Notion’s 200-Billion-Note Machine, Explained in 2 Minutes 👷‍♂️
Every paragraph, checkbox, or emoji in Notion is a “block” row in PostgreSQL. When that single DB hit 20 billion blocks, latency spiked and index bloat set in. Sharding by workspace_id, Notion got 480 logical shards spread across 96 Postgres instances, all routed with a simple hash(workspace_id) % 480. Find the key lessons about choosing boring tech stack and attitude towards observability.
javinpaul
#performance #observability
Inside the 62-Minute 1.1.1.1 Outage 👷‍♂️
There was an incident with Cloudflare DNS last week, but nobody noticed :) The reason for it was A dormant topology rule accidentally grouped Cloudflare’s public DNS prefixes with an internal service. Fixes include dual control on topology edits, synthetic DNS canaries, and diffs that forbid overlap between public and DLS address sets. Textbook example of “latent coupling + manual push”—add it to your outage-drill deck.
Ash Pallarito
#casestudy #pm
The Myth of the “Silent” Lambda Crash 👷‍♂️
A viral gist claimed NodeJS Lambdas die quietly during outbound HTTPS in a VPC. AJ Stuyvenberg though shows that it is rather a feature, not a bug - Lambda was never meant to work as EC2, and the termination happens by design according to scale to zero policy. If you're not happy with this behavior, you have 3 options listed in the article.
AJ Stuyvenberg
#lambda #serverless
The Four Building Blocks of an Agentic App 👷‍♂️
Ever wondered how products like Lovable or Bolt.new work? Beam shares the 4 critical components of such apps: Model Client, Model, Sandboxed Exec and Frontend. Follow the article for details.
#ai #architecture
Time, Angle & Depth — A 3-D Lens on Coupling 👷‍♂️
Another piece on data modelling and context segregation. Discover how how looking at the models from different angles force you creating different classes and apply business logic checks. Starts with a very realistic story :)
#ddd
Big thanks to Nikita, Constantin, Anatoly, Oleksandr, Dima, Pavel B, Pavel, Robert, Roman, Iyri, Andrey, Lidia, Vladimir, August, Roman, Egor, Roman, Evgeniy, Nadia, Daria, Dzmitry, Mikhail, Nikita, Dmytro, Denis and Mikhail for supporting the newsletter on Patreon!