Most backups we audit haven't been tested in months. Some haven't been tested ever. A backup you can't restore is a file. Here are the five things that quietly break and how to catch them before you need them.
The job succeeded. The dashboard is green. The report hits your inbox. And the backup is still broken. Here is how that happens.
One: the agent silently stopped covering a new VM, share, or database. Coverage drift is the most common failure we find. Nobody updated the job when infra changed.
Two: the repository filled up. Retention got quietly truncated to whatever fit. The last good copy is from three weeks ago and you needed yesterday.
Three: credentials rotated. The job ran but wrote nothing, or wrote to a target you no longer control. The logs said 'success' because the runner didn't crash.
Four: the restore path is untested. You have data. You do not have a documented, practiced restore. During an incident is a bad time to learn your tool.
Five: nobody owns it. 'Backups' is on the IT team's plate, but no single person is accountable. Test quarterly. Write down the owner. That's most of the fix.
// questions?
Got a specific situation this article didn't cover? Tell us about it.
Contact us// related reading