A snapshot is a point-in-time copy of your server's disk, taken at the hypervisor. Restoring one returns the disk to exactly that state.
What snapshots are good for
- Taking one before a risky upgrade, so a failed distribution upgrade is a two-minute rollback
- Cloning a configured server as the base for another
- Recovering quickly from a bad configuration change
What they are not a substitute for
Local snapshots live on the same storage as the server. They protect against your mistakes; they do not protect against hardware loss, a compromised hypervisor account, or a legal event that reaches the whole machine. For that you want the off-site option below, or Restic to a second jurisdiction.
Manual snapshots
Every plan includes one free snapshot slot. Take one from the panel or the API:
curl -X POST https://api.nimbusvps.com/v1/servers/SERVER_ID/snapshots \
-H "Authorization: Bearer $NIMBUSVPS_TOKEN" \
-d '{"label":"pre-upgrade"}'Manual snapshots are retained indefinitely. Extra slots are $2/month for five.
Automatic off-site snapshots
The backup add-on takes scheduled snapshots and stores them in a different jurisdiction from the server, encrypted at rest:
- Weekly — 12% of the hardware subtotal, 4 restore points retained
- Daily — 22% of the hardware subtotal, 14 restore points retained
Restoring is one click and takes minutes rather than the hours a full re-upload would.
Consistency, and the caveat that matters
Snapshots are crash-consistent, not application-consistent. A restored snapshot looks to the operating system like a machine that lost power — journalled filesystems recover cleanly, and most databases do too, but a database mid-write can need recovery on start.
For a guaranteed-clean database copy, dump it to disk before the snapshot runs. Belt and braces beats an unpleasant discovery during a restore.