Stop guessing: a practical look at TCP buffer tuning
Everyone copy-pastes the same sysctl block from a 2013 forum post. Here is what those numbers actually mean, and how to pick them for your own links instead of cargo-culting.
Read more →Writing things down so I stop re-debugging them
Field notes from years of running servers, chasing packets, and reading too much kernel documentation.
Why your Docker image is 1.2 GB (and how to fix it)
Image bloat is almost never one big mistake. It is a dozen small ones stacked into layers. A tour through multi-stage builds, layer caching, and the apt-get trap.
Read more →Error handling in Rust without losing your mind
The gap between a toy unwrap() prototype and a service you trust at 3 a.m. is mostly about error types. A pragmatic path from ? to custom errors.
A field guide to writing systemd service units
systemd gets a lot of hate, but a well-written unit file replaces a pile of fragile init scripts. The directives I reach for, and the ones I avoid.
Read more →Reading a traceroute like it owes you money
Asterisks do not always mean packet loss. A short guide to interpreting the noisy middle of a traceroute before you blame your ISP.
Read more →The four shell aliases I install on every box
Small quality-of-life shortcuts that pay for themselves the first week. Nothing clever, just things I got tired of typing out by hand.
Read more →Borrow checker fights I eventually stopped losing
Patterns that turned the borrow checker from an adversary into a slightly pedantic colleague. Cloning is not always defeat.
Read more →