VPS Snapshots Are Not Backups: Know the Difference
Provider snapshots feel like backups but fail the same disasters your server does. Here is the difference and why self-hosting still needs real offsite backups.
A provider snapshot is not a backup, and treating it like one is how people lose everything while believing they were protected. A snapshot is a point-in-time image of your VPS stored by the same provider, on the same platform, tied to the same account. It is genuinely useful for fast rollback, but it fails the exact disasters a real backup exists to survive: the provider going down, your account getting suspended, or a problem that takes out the platform itself. If your only "backup" is a snapshot living next to the server it images, you have redundancy against nothing that matters. I use snapshots, but I never confuse them with backups.
What a snapshot actually is
A snapshot captures the whole disk of your VPS at a moment in time, managed by your hosting provider. Click a button, get an image you can roll the server back to. It is fast, it is convenient, and it is the right tool for a specific job: undoing a change you just made. About to run a risky upgrade? Take a snapshot first, and if it breaks, roll back in minutes. For that, snapshots are excellent.
The convenience is real, so I am not telling you to skip them. I am telling you what they are: a provider-side convenience feature, not an independent copy of your data. The distinction sounds pedantic until the day it is the only thing between you and starting over.
Why a snapshot fails the disasters that count
A backup exists to survive the loss of the primary system. Test a snapshot against that job and it fails, because it shares fate with the thing it is supposed to protect.
- The provider has an outage: your server is down and so are your snapshots, sitting on the same platform.
- Your account gets suspended, even over a billing mistake you did not make: you lose access to the server and the snapshots together.
- The provider has a data-loss event: rare, but it has happened, and it takes the snapshots with the servers.
- You need to move to a different provider entirely: the snapshot is in a format tied to the one you are leaving.
Every one of these is exactly the scenario a backup is for, and the snapshot fails all of them because it never left the provider. This is the same reasoning behind wanting servers that cannot be held hostage: if one party can take away both your server and your recovery, you do not have a recovery.
The 3-2-1 test a snapshot flunks
Hold a snapshot up to the 3-2-1 backup rule and it fails the part that matters most. Three copies? The snapshot is one copy on the provider's storage. Two media? It is the same failure domain as your server. One offsite? No. The snapshot is the opposite of offsite. It is as on-site as it gets, living in the same account, region, and platform as the thing it images.
The offsite copy is the whole point of the rule, and it is precisely what a snapshot cannot be, because a provider snapshot by definition lives with the provider. So a snapshot can be part of your setup, but it can never be your offsite copy, which means it can never be your last line of defense.
What real backups look like alongside snapshots
Use both, for their actual jobs. Keep snapshots for fast, same-provider rollback. Keep real backups for surviving the provider itself.
A real backup leaves the provider entirely: an encrypted, deduplicated copy pushed to object storage you control on a different platform, in a portable format you can restore anywhere. That copy survives your provider going down, your account getting locked, and your decision to move elsewhere, because it does not depend on the provider at all. It is the copy that lets you rebuild on a fresh box with a different company if you have to.
And like every backup, it only counts if you can restore it. A portable offsite copy you have never restored is still just a hope, which is why I run the same restore drills against these that I run against everything. The snapshot you can roll back to in the dashboard; the real backup you prove by rebuilding from scratch.
I use provider snapshots across my portfolio for quick rollback, hosted through HostSSH, and I keep real encrypted backups offsite for everything that would hurt to lose. They are not the same tool and they do not protect against the same things. A snapshot saves you from a bad upgrade. A backup saves you from losing the provider. Know which is which before the day you find out the hard way.