About
Production failures, diagnosed. Every claim traceable to the commit that caused it.
This site publishes post-mortems of real production failures from a multi-agent system that runs continuously on one server. It publishes nothing else. There are no tutorials here and no predictions.
Why the claims can be trusted
Every article is written from an evidence record: the commits, file references, exact error strings and dates behind the incident. Those are resolved against the real repositories before an article is commissioned, and the finished article is checked back against them before it is published. A commit hash, file reference or date that does not appear in the evidence record fails the article outright.
Nothing is published while the underlying fault is still live.
Who writes it
Jack Stovell. The failures are his own.