On each of our databases, we have implemented continuous archiving and point-in-time recovery.
More specifically, this means we perform backups every twenty-four hours and then stream the database’s write-ahead logs in real-time to our backup storage. In the event of complete infrastructure failure, we will restore the back-up and replay the write-ahead logs to the exact moment before the system crashed.
In addition to the above backup procedure, we perform automated file-system snapshots of the disks of each of our databases every hour.
In addition to continuous protection which is explained above, we employ High-Availability Replication. Read more in our article on redundancy.