I write
Incident-style write-ups: one real problem, the guesses that were wrong, the actual cause, the fix, and how to spot it faster next time.
Read the articles →I build backend systems for a living and write down what breaks. This site is where those write-ups live: plain accounts of real problems, for engineers and for anyone curious how software actually fails.
Backend Engineer
Incident-style write-ups: one real problem, the guesses that were wrong, the actual cause, the fix, and how to spot it faster next time.
Read the articles →Backend systems for a living: Node.js and TypeScript services, MongoDB and DocumentDB, AWS, and the Linux boxes underneath them.
See my work →Short versions of the same lessons, plus practical notes on AI tools, on Instagram. Made for people who have five minutes, not fifty.
@theajithakdev →If I have not run into it myself, I do not write about it as if I had. The posts show the dead ends too, because that is where the learning is.
Browse by topic →A sign-in hangs. A query returns numbers that look right and are not. A tunnel stops carrying traffic.
I write down each hypothesis, what it predicted, and what ruled it out. Most of them are wrong.
One observation does not fit. Following it leads to the mechanism, with a command that proves it.
Generalized, stripped of anything private, and published so the next person finds it in a search.
No sponsored posts.
No invented numbers.
No tracking cookies.
I write from real work, so every post is generalized before it goes up. Examples use synthetic data, placeholder domains, and documentation IP ranges.
Questions about a post, a correction, or an idea for one: any of these reach me. The contact page on my main site is the surest.