> ## Documentation Index
> Fetch the complete documentation index at: https://getfloo.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Backups and recovery

> What floo can restore for each managed service, how to restore a Storage object yourself, and how to request a database restore.

floo keeps change history for managed Postgres and 30 days of object versions
for managed Storage. Managed Redis is a cache with no backup. Your source code
lives in GitHub, and resources you connect yourself are backed up by their own
provider.

## What floo can restore

| Resource | What is kept | How to restore |
| - | - | - |
| Managed Postgres | Change history for each database | [Request a restore](#request-a-database-restore); floo runs it |
| Managed Storage | Every overwritten or deleted object version, for 30 days | Yourself, with [`floo storage restore`](#restore-a-storage-object) |
| Managed Redis | Nothing. Treat it as a cache | Not available |
| Source code | GitHub history | Use GitHub |
| Databases, buckets, and APIs you connect with env vars | Whatever their provider keeps | Use that provider |

Each environment has its own data, so a restore names the environment.

## Restore a Storage object

For an object that was overwritten or deleted by mistake:

```bash theme={null}
floo storage versions uploads/report.json --app my-app
floo storage restore uploads/report.json --generation 1700000000000000 --app my-app
```

Add `--env prod` for the prod bucket. The restore copies the chosen version
back to the live path and records an audit event. Versions cover object
overwrites and deletes, not deleting the whole storage service. See
[`floo storage`](https://getfloo.com/docs/cli/storage).

## Request a database restore

The CLI has no database restore command. Email
[team@getfloo.com](mailto:team@getfloo.com) with:

* The app and org.
* The environment: dev, prod, or a preview.
* The resource: Postgres, or Storage if you cannot restore it yourself.
* The table, schema, rows, or object path, if you know them.
* The approximate time before the bad change.
* Whether restoring into a copy instead of in place is acceptable.

floo confirms the restore window, the target, the expected impact, and how to
verify the result before it changes any data.

## Not covered yet

* Database restore from the dashboard or CLI.
* Redis restore.
* Cross-region disaster recovery.
* Recovering a deleted storage service.

## Next steps

> **[Postgres](https://getfloo.com/docs/guides/postgres.md)**
>
> Migrations, schema changes that survive rollback, and connection limits.

> **[Storage](https://getfloo.com/docs/guides/cloud-storage.md)**
>
> Buckets, signed URLs, and object versions.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.