Hope everyone is doing well and staying safe. Welcome to the eleventh edition of this newsletter - here’s what I wrote this week and the absolute best of what I read on the internet.
From the blog
A lot of people responded to my article on using events instead of log for Observability by saying that logs are critical for debugging issues, or that logs needed metrics to complement them, or that the mechanics of logging are different from publishing events. I felt that a follow up was needed for more context and clarifying different terms being used in the discussion. So I wrote one which distinguishes between observability from debugging, logs from metrics, and logging paradigm from logging implementation.
I stand by what I said - events with lots of metadata are the best, most consistent mechanism we can use to build Observability tools. Regardless of the mechanism you use for these events (“log” files, event buses, whatever else - all that is implementation detail), or the aggregates you build them into (e.g. metrics), events are our best option at defining a single source of truth for understanding system behaviour. Even some vendors seem to be coming around, in their own sideways, slippery, vague manner.
What do you think? Article Link
From the great interweb
- Michael Chadwick writes about how his team is using Architecture Decision Records (ADRs) to move fast and not break things as often. Having just joined a startup which unfortunately has little internal documentation, I cannot agree more with the need to document the “why” of technical decisions. Engineers can mostly go through the code to figure out how, but the why is often missed. Use MIchael’s method or any other method, but write down why things are being done a certain way.
- I came across this superb Architecture Playbook discussing various aspects of building a great software system. I’m still going through it so can’t summarize it yet but it is well thought and better written. A definitely recommended read for any software architects/senior engineers. And it’s an open-source book, so you can contribute too!
- Kevin Kwok breaks down Mike Speiser’s style of adding value to his companies. This is a great strategy breakdown for anyone interested in startups and VC thinking. Kevin writes one of the best strategy newsletter I have read on the internet - subscribe if you haven't already.
- I came across this book on town planning and architecture somewhere in Twitterland and for some reason I have been very fascinated with it. A lot of people have compared software engineering and city planning - I hope to learn a bit myself.
That's it for this week folks. Happy weekend!