Who cares about failed backups?

In the past, all too often I’ve worked somewhere, or interacted with an external client that are managing backups. When I ask how they’re monitoring them, or at least receiving notifications on failures, I get looked at like I’m crazy.

To me, from my time in on-premise IT, those were compulsory so you could re-run a backup or investigate why it failed. Just having those notifications saved our bacon with clients multiple times.

I know that AWS is super reliable, and I’ve not seen a time a backup has failed in my time working on AWS. Call me superstitious, but I’m a firm believer in this summary of Murphy’s Law. “Anything that can go wrong, will – at the worst possible moment.”.

Be the change you want to see

So with that in mind I figured instead of screaming into the void, or railing at others for not doing it, I’ll be the change I want to see and provide people with tools to achieve it themselves.

I’ve built a cloudformation template that deploys:

• Backup Vault
• 3 x plans, every 7 days, 14 or 28 days
• backup selection criteria to do it based on tags. Use the key “backup” and values 7, 14 or 28 to use the respective plans.
• SNS topic to subscribe an email address I also constructed it as a terraform module, which also allows the supply of an array of days apart that each plan should execute, and it’ll create them dynamically.

They’re available on GitHub here:

• Cloudformation: https://github.com/joebywan/AWS_coding/tree/main/aws_backup
• Terraform: https://github.com/joebywan/tfmodules/tree/main/aws_backup

So you, yes you! You have no excuse for not having backup notifications on your AWS infrastructure.