LearnHealth & Safety
How To Check If Backups Actually Succeeded
1 min readUpdated 6 Nov 2026
Backups fail quietly: the job finishes, the file exists, and the restore still does not work. Checking coverage and verifying files takes minutes and removes the surprise.
Check coverage, not just the last run
Every database should have a full backup inside your recovery point target, plus log backups if the recovery model demands them. One missing log backup on one database is enough to lose a day of work.
- Full backup age for every database, including new ones nobody added to the job.
- Log backup recency for databases in the full or bulk-logged recovery model.
- Databases whose recovery model changed without a matching backup change.
Verify the file, then test the restore
- 1.Run a restore verification on the latest backup file, not just the job output.
- 2.Confirm the file exists on the path the job reported, with the size you expect.
- 3.Restore a real database to a scratch server on a schedule, and time it.
- 4.Record the restore duration — it is your actual recovery time, not your estimate.
The quiet failure modes
- A new database that no job covers.
- A backup written to a disk that filled up afterwards.
- A log backup chain broken by a recovery-model change.
- A backup that succeeds but takes so long it overlaps the next one.
Recovery is measured in restored data and elapsed time. Everything else is paperwork.
What to remember
- Check backup age per database, not per job.
- Verify the file, do not trust the job status.
- Restore for real, on a schedule, and time it.
- Watch for new databases and broken log chains.
