It will be a good idea, that we take backups of the sqlite databases in the directory ( where they exist ) when pegasus-db-admin tries to update the database.

this will be helpful for users when something goes wrong with the pegasus-db-admin tool. updating the existing database for example the master workflow db , makes it hard to figure out what the bug was when things go wrong.

pegasus-monitord does something similar in the replay mode. takes back up and calls the backup .00X , where x is an increasing number dependent on the number of backups taken.