A place to write down things I had to learn the hard way.
I work on infrastructure: storage systems, network plumbing, and the operational side of keeping them predictable. This site collects notes on failure modes that are easy to misdiagnose, measurements that are easy to compute wrong, and assumptions that hold until they do not.
Everything here is written from production experience rather than specification reading. Where I am unsure, I say so. Where a claim is narrower than it sounds, I try to draw the boundary.
Corrections and counterexamples are welcome. If you have run into the same problem and solved it differently, I would like to hear about it.