Backups in the dashboard.

Inspect your instance's recovery points and tail-shipping status. You'll need a running instance and the Backups area is read-only — there are no destructive actions here. Restoration lives on its own page; see Restoration.

1. 1. Open the sidebar and click Backups (or jump to /app/backups).
2. 2. Pick your instance from the Instance dropdown.
3. 3. The Recent snapshots table lists per-tier EBS snapshots (Whisper 7d, Thunder 90d, Storm 365d).
4. 4. The Sealed-segment shipping table lists every WAL segment + checkpoint shipped to S3.
5. 5. The Continuous tail-shipper panel shows last-shipped timestamp, last-shipped LSN, and effective RPO. Pause/resume control is available when the Intra-Segment PITR add-on is active.

What each piece means

Sealed-segment snapshots

Taken automatically: daily on Whisper, hourly on Thunder, every 15 minutes on Storm. Each snapshot is an EBS-backed image of the sealed WAL segments plus the on-disk indexes at that point. Restorable to a fresh instance via the dashboard or HTTP API.

Sealed-segment shipping (GA, all tiers)

Once a WAL segment is sealed, it ships to S3 continuously. Restoration replays segments up to a chosen target timestamp at sealed-segment precision, which is sufficient for most workloads.

Continuous tail-shipper (Intra-Segment PITR add-on)

Sub-second LSN-precise PITR. WAL frame v2 embeds timestamp_micros at write time; the tail-shipper pushes frames between seal events on a sub-second cadence so the archive is never more than a few seconds behind production. Tune the window/min-delta to trade S3 PUT cost against RPO precision.

RPO (Recovery Point Objective)

The maximum data loss window if the writer dies right now. With sealed-segment shipping the RPO is the time since the last seal event (typically a few minutes). With the tail-shipper running, the RPO is the configured window — typically under one second.

Common errors

• backups add-on not enabled on the tail-shipper section — enable Intra-Segment PITR from /app/billing/addons.
• no snapshots yet — the instance is younger than its tier's snapshot cadence. Wait for the next scheduled snapshot.
• tail-shipping disabled — the add-on is enabled but the operator-side spawn flag has not been set. Contact support@originchain.ai.
• RPO drifting — last-shipped timestamp is more than a minute behind wall clock. Check /app/metrics for shipper backpressure or S3 throttling.
• instance paused — snapshots and tail-shipping pause when the instance is paused. Resume from /app/instances.