Feedback for customer:

To restore the indexes, and jira-administrator, which is supposed to be up 24/7, had to login to jira via web AND SSH in order to find out the filename of the last know backup and to trigger a restore index, and one minute later to trigger a background reindex.

For such a feature to be effective the entire process must be automated. Once JIRA detects that the indexes are broken, it should try to restore last valid backup and do a background reindex after this succeeds.

If it can even send a message when this starts, ends is even better, but that would be an enhancement over the basic index recovery option.

Also, please note that I do find extremely counter-intuitive the fact that the restore index asks you about a filename instead of just showing you the latest available one (or a list of them.). This slows you down considerably as you need shell access to the machine to find-out the name of the archive.