Why I stopped chasing the fastest link
Latency numbers look great on a dashboard and tell you almost nothing about a bad evening. A notebook of small measurements, taken at the same hour for a month, beats any single benchmark.
A tiny checklist before touching a server
Snapshot first, read the logs second, change one thing third. Most outages I have seen were caused by a second change made before the first one was understood.
Plain text is a feature
Everything in this notebook is a static file. No scripts, no trackers, nothing to patch. When something is boring, it keeps working.