The suggestion is to have a way to configure the Integrity Checker in such a way, that it would list all the detected inconsistencies without automatically fixing them.
The customer shared that at present, in order to achieve this, they have to use a very tedious workaround, where they take a snapshot of existing data (NFS + DB), run Bitbucket with the Consistency Checker, collect its output, then restore the data from the snapshot and manually fix the inconsistencies that were being reported. This is a very lengthy and tedious method, extending Bitbucket's downtime, and potentially losing any changes that may be received before the data is restored from a snapshot.
Why the customer wants to disable the automatic reparation of inconsistencies?
Since consistency checker fixes a few inconsistencies with a user name as "Integrity checker", the customer is not sure for example a merge being done which suppose to trigger a CI/CD build and it may not trigger a build and it is always a better situation to have the option to fix the inconsistencies or not. ( this will add more flexibility during disaster recovery situations)