Architecture Weekly Issue #142. 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.
Saving 80% of cloud cost by moving from AWS to Cloudlfare 👷♂️
Baselime is an observability solution recently acquired by Cloudflare. With this, it was a natural choice to transition the payloads from Lambdas to AWS workers. But you never gonna believe the cloud cost plummeted 80%. Find out architecture before and after the transition and multiple tech details inside the article.
Boris Tane
#serverless
Why do I need CDC? 👷♂️
Do you know what Change Data Capture? Do you know about Debezium and other solutions? Yeah, but do you know why you need it in the first place? Grab a refresher on the reasons why CDC is required.
ByRobin Moffatt
#db
Fault Injection Service for AWS Lambda 🍼
Chaos Testing was invented at Netflix, but quickly got adopted in every mature software organization. Chaos testing for serverless though was underdeveloped, but not anymore! Now with Fault Injection Service coming to lambda, you can find out how your system behaves if something breaks there.
#serverless #reliability
Follow-Up
Scanning documents with Claude 3 Sonnet and serverless 👷♂️
Do you remember ABBYY was firing engineers recently? No wonder, as multi-model LLMs are now capable of doing the job faster, better and cheaper. At supplied.eu, we moved from Abbyy Vantage solution to ChatGPT API, and we run it in cloud. Here you will find an article of how to do the same, but with Claude. I think that having a queue and a second lambda is an overcomplication here, but can be considered as well.
#llm #serverless
Vector Databases Are the Wrong Abstraction 👷♂️
If you're building an application with RAG, if probably faced a problem of synchronization between the source data and the embeddings in the vector database. This articles argues, that the problem in the wrong abstration: the embeddings are derived data and should be stored right next to the original one, essentially making them like indexes. Find out the solution to this problem:
Matvey Arye
#db #ai
Decision-Making Pitfalls for Technical Leaders 🍼
I frequently observe engineers making suboptimal technical decisions because multiple reasons: focusing too much on the former experience, rather than the problem, failing to account for risks and others. That's why I recommend to read this article: it does a good job how to think for better decision making.
View all posts by Chelsea
Replication in Distributed Systems - Part 1 🍼
Once you need to improve performance after data exceeds one-machine capacity, you typically look at the replication. In this concise article you will find explanation how it happens and what are the different types of replication mechanisms exist.
Manthan Gupta
#distributedsystems
The Engineering behind Booking.com's Ranking Platform 👷♂️
What happens when you type "Hotel in Tallinn" in booking.com? Surprisingly, there's a lot of stuff from going from microfrontends to API Gateway to complex ML infrastructure, that should account for your personal preference. Booking.com explains the components and connections of the search at global scale.
#casestudy
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!