we proved it rebuilds by destroying it

A rebuild you have never performed is a hypothesis. The proof is a reset cycle: stop every workload, remove it, resubmit from the repository, and check that the services come back and answer. Running that deliberately, on a quiet day, converted a hope into a routine and made the phrase "we can rebuild it" something with a receipt.

It also exposed how much of each machine lived only in someone's memory. Firewall rules, service definitions, the order in which things had to start — all of it had been typed once and remembered since. Those facts moved into the repository, machine by machine, each one a directory with its shape written down and its exceptions admitted.

The cost is writing, and it is not small. A repository that can rebuild a machine has to be maintained as if it were the machine, and the temptation to make one quick change by hand never goes away. The rule that keeps it honest is that nothing counts as done until it has been done from the repository, on a machine that had nothing on it.

What this proves is that reproducibility is a property you demonstrate by destroying something, not a claim you put in a document. The habit is expensive, and it is the difference between a lab and a set of servers nobody would dare turn off.