Back to Resources
Guide

5 Backup Mistakes Our Monitoring Is Built to Catch

These five mistakes aren't hypothetical; they are the failure patterns our monitoring agents are built to detect in SME environments: silent job failures, shrinking backup sizes, untested restores, a single copy, and no offsite protection.

These five mistakes aren't hypothetical; they are the failure patterns our monitoring agents are built to detect in SME environments: silent job failures, shrinking backup sizes, untested restores, a single copy, and no offsite protection. Each one below comes with a five-minute check you can run today.

We build backup verification software, which gives us an unusual vantage point: we designed detectors for the ways backups actually fail, not the ways people worry they'll fail. The list that emerged is short, unglamorous, and devastating: every entry below has quietly ended businesses.

1. The silent failure: the job that stopped and nobody noticed

The most common mistake isn't a bad backup; it's a backup that stopped happening. A password changed, a disk filled, a service didn't restart after an update, and the job that ran faithfully for two years quietly stops. Nothing turns red, because the software that would report the failure is the software that stopped running. The gap is discovered months later, on the worst possible day.

Five-minute check: open your backup destination (not the backup software, the actual folder or bucket) and look at the newest file's date for each system. If anything surprises you, you've found a silent failure.

What our monitoring does: freshness is checked from the outside, against the files themselves: every target has a maximum age, and a backup that stops arriving turns stale and raises an alert whether or not the job's own software has an opinion.

2. The shrinking backup: an early ransomware tell

Backup sizes are boring, until they change. A backup that suddenly halves usually means part of the source vanished or stopped being included. A backup whose contents suddenly can't be compressed or whose size pattern shifts oddly can mean something worse: encrypted files replacing real ones, which is what ransomware looks like from the backup's point of view, sometimes before anyone notices on the live systems.

Five-minute check: list your last 10 backups for the main system, and just read the sizes. You are looking for cliffs: a sudden drop or jump with no explanation (a big cleanup, a new dataset).

What our monitoring does: size history is tracked per target with abnormal-growth and abnormal-shrinkage detection, and restored data is scanned for ransomware indicators: entropy spikes, mass file-extension changes, ransom-note patterns. A hit quarantines the affected verification until an admin reviews it.

3. The untested restore: backed up for years, restored never

The uncomfortable industry secret: many backups don't restore. Corrupted archives, missing dependencies, a database backup taken in a mode that can't actually be recovered, an encryption key nobody saved. None of that shows in a job log; a backup can report success every night for years and still be unrestorable, and there is exactly one way to find out. On purpose now, or by force during a disaster.

Five-minute check (to schedule, not perform): put a one-hour restore test in the calendar this month: restore one system to a spare machine or VM, open the result, write down the date and how long it took. That record is worth more in an audit than a year of green logs, as auditors will confirm.

What our monitoring does: restore testing is the core of the product: scheduled real restores of databases (SQL Server, MySQL, PostgreSQL), files and Hyper-V/VMware VMs, with the restored data validated against expectations and the duration measured, so your recovery time is a number, not a guess.

4. One copy: backup and original sharing a fate

A backup on the same disk, same server, or same RAID array as the original isn't a backup; it's a second casualty. The same applies logically: if one administrator account can delete both the data and its backup, ransomware that captures that account gets both, which is precisely what modern ransomware operators do first.

Five-minute check: ask "what single event destroys both the original and the newest backup?" If any answer exists (one dead server, one stolen credential, one flood), you have one copy, whatever the job log says.

What our monitoring does: backup chains are tracked against the classic 3-2-1 rule (three copies, two media, one offsite), and where your storage platform supports immutability (object lock), we verify your copies really are locked: an attacker with credentials still can't rewrite them.

5. Nothing offsite: fire, theft and encryption reach the same room

The office fire, the office burglary and the office ransomware outbreak share a property: they happen where your servers are. A backup NAS in the same room protects you from a dead disk and nothing else. The old objection (offsite was expensive and slow) hasn't been true for years; object storage costs a few ringgit per hundred gigabytes per month.

Five-minute check: point at the physical location of your most recent backup copy that is not in your building. If you can't, this is your gap, and it compounds every other mistake on this list.

What our monitoring does: offsite copies in your own cloud storage (S3, Azure, Google Cloud, Cloudflare R2, Backblaze B2, Synology, and S3-compatibles) are first-class verification targets: we don't just check they exist, we download and restore-test them, because an offsite copy nobody has ever read back is mistake #3 wearing mistake #5's coat.

The pattern behind all five

Every mistake on this list survives on the same nutrient: nobody is looking. Each one is invisible in the daily job log and obvious the moment something independently checks the backups themselves: their freshness, their size, their restorability, their separation. That independent checking is the entire idea of Backup Governor; the five-minute checks above are the manual version, and they're genuinely worth running today. For turning the results into something you can show your boss or auditor, see How to Prove Your Backups Actually Work.

Free for one node

Stop being the one who has to notice

Freshness, size anomalies, ransomware indicators, restore tests and offsite verification, watched continuously, free for one node.

Get started free
No credit card requiredCancel anytime

Want to see these best practices in action?

Start for free and experience proper backup monitoring firsthand.

No credit card required
Cancel anytime