Diagnostics, reports, and backups

Diagnostics

Diagnostics provides separate paginated views for recent sync jobs, failed jobs, and failed documents. Each table has its own connector, status, or text filters.

Available actions include:

• requeueing a failed job,
• retrying extraction for an individual or selected document,
• selectively reindexing document IDs,
• running consistency repair for selected content,
• opening the source when a provider URL is available,
• exporting diagnostic data for support or analysis.

Prefer selective retry or document reindex over a full rebuild.

Audit log

The Audit log resolves actors and actions into readable labels while retaining raw evidence in the event detail. Search locally, filter by decision and date range, then confirm a scoped server-side CSV export. Access to reading and exporting audit evidence can be assigned separately.

Backup and restore

Backup archives can include the database state, original and derived artifacts, connector configuration, permission snapshots, and index data needed to restore an installation.

1. Store the generated encryption token in a password manager.
2. Select Download backup for an encrypted portable archive, or create a local snapshot.
3. To restore, select an archive, enter the matching token, and confirm the destructive action.
4. Choose whether to retain restored indexes or rebuild all indexes.

Keeping restored indexes avoids repeating potentially expensive embedding and model work. Choose full reindex only when index artifacts are unavailable, incompatible, or intentionally excluded.

Automated snapshots may be available for your installation. Always test restore on a separate environment before relying on a backup policy.