A green backup dashboard only proves that the scheduled job reported success.
The harder question is whether the data can be restored when the system has changed. I have seen a backup file exist exactly where it should be, while the restore still failed because the schema had moved on. In another case, the problem was not the file at all. An IAM permission had been removed during a rotation and nobody linked that change back to the backup account.
That is why I like a small restore test, not just a backup status check. Restore a slice into a test instance, record the timestamp and confirm the row count or service output. The result should land somewhere visible so the team can see the test, not only the storage job.
The runbook needs the exact role, test instance and location of the keys. A new engineer should be able to run the restore without guessing which account matters.
Backups are only comforting when restore paths are tested. The green tick can stay, but it should come after the restore proves the backup is usable.