Backup, restore, and migration strategy

## Summary
Define and implement backup, restore, and migration workflows for self-hosted RFChat deployments.

## Scope
- Define what must be backed up, including database state, files, and critical configuration.
- Provide a documented restore flow that can rebuild a working deployment.
- Cover schema and storage migrations across versions.
- Make backup/restore expectations explicit for operators before data-bearing features expand.

## Acceptance Criteria
- [ ] Operators have a documented and tested path to back up all durable user data.
- [ ] Operators can restore into a working deployment with messages, media, and core settings intact.
- [ ] Version upgrades have a documented migration story for schema and storage changes.
- [ ] Failure modes and unsupported downgrade/restore cases are made explicit.

## Operational Notes
- Self-hosted trust depends on this being boring and recoverable.
- Restore behavior should not rely on hidden cloud-only assumptions.