Files
Whole folders or chosen files, back in place or to another path or server.
How restore works
What you can restore
Section titled “What you can restore”Volume
A Docker or Podman volume, the original or a new one.
Container or stack
A container’s filesystem or a compose project.
Database
Into the original instance, under another name, another instance, or a new container of the dump’s version.
Download only
Get the files without writing to any server.
Who does what
Section titled “Who does what”| Step | Who | Permission |
|---|---|---|
| Request a restore | Restore Operator or Administrator | Restore › Request |
| Approve or reject | Administrator or a custom approver role - never the requester | Restore › Approve |
| Execute an approved restore | Restore Operator or Administrator | Restore › Execute |
Safety nets
Section titled “Safety nets”Safety snapshot
Before overwriting, the target is backed up and kept 14 days. If that fails, the restore stops.
Roll back
A failed or cancelled restore offers Roll back from safety snapshot - a new request with the usual approvals.
Validation
Database restores run the engine’s checks and compare row counts; your read-only queries run after them.
Approvals expire
An approval older than 24 hours sends the request back for approval.