Best practices for Cloud backups

Continuing our discussion on how to backup databases to the cloud here are a few official and unofficial best practices.

  • Use unique filename; better yet add timestamps to the backup to make sure the cloud blob storage doesn’t overwrite the existing file. Since the whole thing happens outside your premise there is no File overwrite warning.
  • Set blob Storage to private to prevent unwanted access or better yet make sure all connections between on premise and cloud happens using AD and VPN.
  • Try to make sure when using SQL instance such as one hosted on VM the blob storage is in the same region as the VM to avoid data transfer cost between MS datacentres.
  • Clean up old backup files from the cloud periodically, some useful tools include Azure storage explorer mentioned in a previous blog.
  • Use Compression where possible to allow better data transfer.

Always keep in mind that network latency will become a factor when performing cloud backups ideally for mission critical databases it makes sense to still continue using on premise approach, however for small projects or database already hosted on the cloud it makes more practical sense to keep everything on the cloud allow infra teams to focus on the hardware needed for the core business.