Periodic Historical Archiving
Periodic archiving continuously saves data to a local directory or external drive on a schedule, supporting long-term retention, offline analysis, and data handover.
Configure each plan independently
Section titled “Configure each plan independently”Create multiple plans with separate directories, scopes, and intervals. For example, one can save historical values for trends while another saves business records and configuration snapshots.
| Setting | Description |
|---|---|
| Interval | 1–30 local calendar days |
| Time zone and first cutoff | Define consecutive archive windows; the first cutoff should be in the future |
| Historical scope | Select all or specified acquisition, aggregation, and shared history points |
| Business categories | Select records, configuration, and relationship data by category, or select all business tables |
| Runtime files | Optionally include ems.yaml, CommBridge configuration files, and display files when present |
| Storage target | Existing local directory, or mounted external drive with verified identity; plans can use different disks |
Local directories are convenient. Use independent storage to isolate whole-device or same-disk failures. External-drive mode verifies mounting and disk identity to prevent accidentally writing to the system disk when the external drive is unmounted.
From creation to completion
Section titled “From creation to completion”- Select the scope, target, and interval, then save the plan.
- Check the storage target and retention period before enabling the plan.
- Wait for the archive window to close, then check execution status, packages, and errors.
- The window is archived only after its manifest, verification results, and completion marker are confirmed.
Successfully saving or enabling a plan does not mean initial protection is complete. Archiving advances through consecutive windows. A disconnected disk, insufficient space, or failed verification prevents the completion position from advancing. Once conditions recover, the same window can continue or be retried.
Files and resource control
Section titled “Files and resource control”Historical values are stored as CSV and other business tables as JSONL, with ZIP packages, SHA256 checksums, and an overall manifest. Archiving runs in segments according to execution windows and resource budgets, sharing query resources with manual exports. Manual exports can take priority at segment boundaries.
The resource page warns about incomplete initial archiving, incomplete scope, retention gaps, and target-drive problems. Archive intervals must match retention periods: if data is cleaned up before archiving, retries cannot recover it.
The current archive feature has no online restore API and does not guarantee a consistent point-in-time snapshot across multiple tables. Archive files cannot be uploaded directly as configuration backup/restore packages.
APIs and next steps
Section titled “APIs and next steps”Use /ems-data/archive/plans to manage multiple plans. Categories, status, and run records use /categories, /status, and /runs, all under /ems-data/archive. Updates and enable/disable requests require the latest revision to avoid overwriting another administrator’s changes.
Review Storage capacity and retention policies before configuration. For a one-time file, use Historical export. See the archiving Skill for detailed invocation steps.