Short notes on running small services close to the network edge.
What this is
A small collection of operational notes: how to keep a single-box service observable,
how to reason about latency budgets, and what actually breaks first under load.
Recent entries
Measuring real RTT instead of trusting a single ping sample.
Why connection reuse matters more than raw throughput for API traffic.
TLS session resumption: what the handshake log can tell you.
Keeping a small Postgres instance healthy on a 1 GB box.
Elsewhere
These notes are published as-is. No comments, no newsletter, no tracking.