Snapshots that run on a schedule, not on memory.
Attach a policy once — hourly, daily, weekly or monthly with a retention count — and every volume protects itself. Copies land on separate secondary storage inside our Indian facilities at ₹2.75/GB, 42% below EBS snapshot pricing in Mumbai.
₹2.75/GB-moPolicies hourly → monthlyRetention per policyRestore one-clickResidency India onlyFour schedules, stacked to taste — all in IST
Policies stack on one volume: a common production pattern is hourly×24 + daily×7 + weekly×4 + monthly×12.
| Schedule | Runs | Typical retention | Protects against |
|---|---|---|---|
| Hourly | at :MM every hour | keep last 24 | Fat-finger deletes, bad deploys within the day |
| Daily | HH:MM IST daily | keep last 7 | Yesterday's mistake discovered today |
| Weekly | chosen weekday | keep last 4 | Slow corruption noticed late |
| Monthly | chosen date | keep last 12 | Compliance and audit look-backs |
Honest note on consistency: snapshots of a running VM are crash-consistent — like the state after a power cut. Filesystems handle this fine; busy transactional databases may not. For MySQL/PostgreSQL under write load, quiesce first (fsfreeze or a pre-snapshot dump) or add agent-based backup. We'd rather tell you this now than after a restore.
Built for the 2 a.m. restore, not the brochure
Set-and-forget policies
Hourly to monthly schedules with per-policy retention counts. Old snapshots age out automatically — no lifecycle scripts to babysit.
Instant manual snapshots
One click before every risky migration or schema change, taken while the VM keeps running.
VM snapshots too
Capture a whole VM — disk plus memory — as a point-in-time you can revert to before patching or upgrades.
Restore or clone
Roll a volume back in place, or spin the snapshot into a fresh volume and mount it alongside — compare before you commit.
Off-host by design
Snapshots copy to separate secondary storage — a host or primary-storage failure can't take your restore points with it.
Full audit trail
Every snapshot, restore and deletion is an event with an actor and timestamp — the record your security review asks for.
CloudStack snapshot policies doing the remembering
Each schedule is a CloudStack snapshot policy — interval type, time in IST, and a max snapshots retention count. The scheduler fires, the hypervisor snapshots the qcow2 volume, and the copy streams to secondary storage.
Restores create a volume from any snapshot via one API call — the same call the console button makes.
- Retention enforced per policy — oldest snapshot pruned automatically
- Copies on secondary storage, physically separate from primary NVMe
- Both copies inside Indian facilities — DPDP-clean by construction
- API, CLI and Terraform parity with everything in the console
- Events feed for SIEM ingestion