A dry run mode of time-series remediation scripts can show what the remediation script will do, without actually modifying the underlying buckets.

Some of the things that this dry-run mode can report is this is what the current (corrupted) bucket contains and its metadata, this is what the new bucket we would make would make and its metadata, and highlight the differences.

This could be helpful for TSEs running the script to explain what would happen as a result of the script.

This is an intermediate way of addressing the idea that we should have specific invariants that match between the two buckets (the corrupted bucket and the new corrected bucket based on the previous corrupted bucket), but rather than guaranteeing them, we highlight the differences because we can't guarantee the invariants match with the previous bucket due to its corrupted state.