I have spent the better part of a decade keeping other people's services online: tuning kernels, untangling networks, and explaining to managers why "just restart it" is not a long-term strategy. These days I work on infrastructure tooling for a mid-sized SaaS company, mostly in Rust and Go, with more YAML than I would like to admit.
This blog is my external memory. Every post here exists because I solved the same problem twice and got annoyed enough to write it down. If something I wrote saves you an afternoon, that is the entire point.
What I tend to write about
- Linux internals — the parts of the system that only become interesting when they break.
- Networking — TCP, routing, latency, and the long tail of "it works on my machine."
- Containers — building lean images and running them without surprises.
- Rust — because I keep reaching for it and keep learning something new.
A few things I believe
The best monitoring is the alert that never fires because you fixed the cause. The second best is the alert that tells you exactly what to do.
I favor boring technology, plain text configuration, and tools that fail loudly. I distrust anything that "just works" until I understand why it works. Most of my worst outages came from a clever abstraction I did not fully understand.
Elsewhere & contact
I do not run comments here — they are more spam than signal. If you want to reach me, I read mail at mara [at] notesonsystems [dot] dev and usually reply within a week. Corrections are especially welcome; I would rather be right than look right.
This site is intentionally plain: static HTML, a single stylesheet, no trackers, no cookies, no JavaScript. It loads fast and it will still load in ten years. That is a feature.