Most engineering writing is about building things. This is about the moment something built breaks in production, and about the habit of reading a system’s behaviour before reaching for a fix.

Each incident is presented the way it actually happens: a symptom, the constraints that narrow it, the evidence, and then the question. The diagnosis is withheld until you ask for it, because the part worth practising is the part before the answer.

Nothing here is a puzzle with a hidden trick. Every incident is something that genuinely happens, and every fix is one that has genuinely worked.