What this is

Teximos publishes writeups of production failures in working software. One incident per piece, with the real numbers, the code that caused it and what the fix actually changed.

The bias is toward the failures that look boring and are not. A pricing table with two rows missing. A guard pointed at the wrong set. A transaction that protected the write and not the decision in front of it. These are rarely written about, because they are embarrassing and because they are specific, and the specifics are exactly what makes them useful.

Nothing here is a tutorial, a framework comparison or a list. If a piece cannot name what broke and show the measurement, it does not get published.

Who writes it

Jeremiah Diana. The incidents published so far are ones I caused and then fixed in a product of my own, which is the only reason the logs, the numbers and the offending code can be shown in full rather than paraphrased.

Corrections are welcome. If something here is wrong, it is worth fixing in place rather than leaving.