Preserving history

Two situations sit next to migrating in: moving between two Cobblr workspaces, and keeping a copy of what you had before an import so you can go back.

Moving between two Cobblr workspaces

If both sides are Cobblr, you have better tools than CSV.

• A backup is a full point-in-time copy of a workspace: its setup, every entity row, and every uploaded file, in one zip. Restore it into a fresh workspace and you get the same workspace back, content and all. A backup only restores into an empty workspace; it will not merge into one that already has data.
• A blueprint is the shape without the data: the modules a workspace enables, its bundles, saved views, public surfaces, and dashboard layout. Install a blueprint and you get an empty workspace that looks and behaves the same, ready for its own contents. Two people can install one blueprint and share nothing but the structure.

Restore keeps references intact. Inside a backup the original ids are preserved, so links between rows (an item's location, a photo on a record) are remapped consistently when the workspace is rebuilt. A backup is a zip of JSON and raw files, so even with no Cobblr in front of you, you can open it and read your data.

See Backup and export for the buttons and what each artifact carries.

Keeping a copy before you import

Before a large import, take a backup of the target workspace. If the import lands wrong, restore the backup into a fresh workspace and try again against a clean copy. An operator-level copy of the whole box is the other half of this: see Backups and restore tests for the database dump and the bind-mounted data tree that back the entire instance.

What a backup does not carry

A backup reproduces what you manage and what you uploaded, not your credentials. It leaves out API tokens, workspace memberships and invitations, and channel secrets like a Discord webhook URL or SMTP password. After a restore you re-plug those. That is deliberate: an archive should never resurrect a live credential.