​
Overview

A snapshot policy takes snapshots of a volume on a schedule and keeps a fixed number of them. It replaces remembering to do it manually, and — unlike manual snapshots — it prunes.
Navigation: Storage → Snapshot Policies. Titled Snapshot Policies, subtitled “Manage automated snapshot scheduling policies”.

​
Two ways in

Same schedule model either way. The per-volume route is quicker when you’re already looking at the volume.

​
Columns

Policy ID is sortable. Expanding a row repeats the fields in a readable layout.
The list identifies volumes by ID, not name, so matching a policy to a volume means comparing identifiers. Coming from the volume’s own Recurring Snapshots dialog avoids that entirely.

​
Creating a policy

1

Choose the volume

Virtual Machine first, then Volume — the volume list is filtered to that machine’s volumes in the Ready state. A volume that isn’t Ready can’t be given a policy.
2

Pick the interval

Interval defaults to Daily. The time fields that appear depend on your choice.
3

Set retention

Max Snapshots to Keep — between 1 and 10. Outside that range the field reports Must be 1–10.
4

Set the time and timezone

The fields vary by interval (below). Timezone offers UTC plus nine common zones.
5

Check the preview

Schedule Preview states the schedule in words before you commit — read it, since an off-by-one on day-of-week is easy and invisible in the raw fields.

​
Fields by interval

Available timezones: UTC, America/New_York, America/Chicago, America/Denver, America/Los_Angeles, Europe/London, Europe/Paris, Asia/Tokyo, Asia/Shanghai, Australia/Sydney.
Day of Month goes up to 31. A policy set to day 31 will not run in February, April, June, September, or November. For a genuinely monthly snapshot, use day 28 or lower.
Choose a timezone deliberately rather than accepting a default. A policy on Europe/London shifts by an hour twice a year relative to UTC; UTC doesn’t drift, which makes correlating snapshots against logs much easier later.

​
Retention, and what it does not cover

Max Snapshots to Keep caps how many snapshots this policy retains. As it takes the eleventh, the oldest is pruned.
Two things retention does not do:It doesn’t cover manual snapshots. Snapshots you took by hand are outside the policy’s accounting and are never pruned automatically.It doesn’t survive deleting the policy. The delete dialog says so: “No new snapshots will be created automatically. Existing snapshots will not be affected.” Retire a policy and its accumulated snapshots stay, consuming quota forever, until you delete them yourself from Snapshots.

​
Editing and deleting

Edit Policy changes the schedule, retention, and timezone. The volume is fixed — shown read-only — so moving a policy to a different volume means creating a new one. Delete Policy stops future snapshots and leaves existing ones alone.
Retention is capped at 10 per policy. If you need a longer history than 10 intervals, run two policies on the same volume at different intervals — a daily policy keeping 7 plus a weekly keeping 4 gives you a week of fine granularity and a month of coarse coverage, for 11 snapshots.

​
Worked example: tiered protection for a database volume

The goal: recover from this morning’s mistake, and from a corruption noticed three weeks late.
1

Daily policy for recent recovery

Interval Daily, Hour 02, Minute 00, Timezone UTC, Max Snapshots to Keep 7.Preview should read Every day at 02:00. That’s a week of daily restore points.
2

Weekly policy for the longer tail

A second policy on the same volume: Weekly, Day of Week Sunday, Hour 03, Minute 00, Max Snapshots to Keep 4.Offset by an hour from the daily so the two don’t fire together. That’s a month of weekly restore points.
3

Verify both are listed

The Snapshot Policies list should show two rows for the same Volume ID, one Daily and one Weekly.
4

Check the output after a week

In Snapshots, filter Type to Daily — you should see 7 and no more. That confirms pruning is working.
Snapshots are not a substitute for backups. They live in the same storage as the volume they protect, so they help against mistakes but not against losing the underlying volume. Pair a policy with a backup plan for anything that matters.

​
Empty and error states

Snapshots

The output, and how to tell scheduled from manual.

Volumes

Attach a policy from the volume’s row menu.

Backup Schedules

The equivalent for whole-machine backups.