We rely heavily on Amazon RDS type services to maintain good, point-in-time restoration. I don't think we ever get more than 15 minutes ahead of the last restore point (MS SQL, PostgreSQL and MySQL/Aurora).
Daily in the early morning hours and we have a read replica which syncs instantaneously when changes are applied to the master database.
We automatically backup our SQL DB every day, and keep backups for a month.
NoSQL databases are also backed up, but as it's the aggregated data we create from this that is important (which is saved to SQL and generated pretty much instantly), these are kept only for a week.
Both types of databases are replicated to multiple nodes, so unless an attack is in place that deliberately deletes data (deletion is also replicated), or we mistype a query and write DELETE instead of SELECT, our data is safe. Still, better to be safe than sorry.
we automated backing up production db and we do it everyday. We will erase backups which are older than 30 days.
j
stuff ;)
as all the others daily, but we compress and encrypt our databases as well so they cannot be stolen.