Disks on the same host, and a week you can go back to.
The SSD in your shape is local NVMe on the hypervisor running the instance, not a volume reached over the network. Everything you add to it — extra volumes, snapshots, a rolling backup — is priced the same way as the rest: a line you can see, on an invoice you can check.
Figures here are demo data. The retention window, the rate and the arithmetic are the real ones.
Four things, and they do not pretend to be each other.
A root disk, extra volumes, snapshots you take, and a backup policy that runs whether you remember it or not. Each is a separate object with its own lifecycle, because merging them is how people lose data believing they had a copy.
The root disk — local NVMe
The SSD number in your shape is cut from NVMe attached to the same hypervisor the instance runs on, metered per GB per month. There is no storage network between the process and its disk, which is the difference you feel on a database rather than in a benchmark.
Additional volumes
Create a disk independently of the instance, attach it, grow it, detach it and move it to another machine. A quote endpoint prices a volume before you create it, so you see the figure for the volume you are about to make rather than working it out from a table.
Snapshots, on demand
A point-in-time copy of a volume, taken from the instance page in seconds — the thing you do before a migration you are not sure about. Revert the volume onto it afterwards, cut a brand-new volume from it, or delete it when the change stuck.
The 7-day rolling backup
A daily snapshot policy with a week of restore points, provisioned with the instance and enforced at the hypervisor rather than by an agent inside your machine. It is priced at 20 % of the instance per month, and it is opt-in.
Grow the disk without taking the machine down.
A volume is its own object. It outlives the instance it is attached to, it can be grown in place, and it can be detached and re-attached somewhere else — which is what makes a restore into a clean machine a routine operation instead of a project.
Expansion without downtime
Growing an attached volume does not require detaching it, stopping the instance or rebuilding anything. You ask for the larger size, the platform prices it first, and the disk is bigger — the filesystem inside it is yours to extend, as it should be on a machine you have root on.
Restore into a clean machine
A snapshot can cut a brand-new volume, and that volume attaches to a freshly deployed instance. That is the safe shape of a restore: the suspect machine stays exactly as it is for you to look at, and the recovered data comes up beside it rather than on top of it.
No egress line on the invoice
Data leaving your instances is not metered and there is no bandwidth allowance to exceed. Pulling a 400 GB database dump out on the day you decide to leave costs nothing — which is the only moment that promise is worth anything.
What it is, and what it is not.
It is a daily snapshot policy with seven rolling restore points, held at the hypervisor. It is not a guarantee that any particular byte comes back, and our own SLA says so — you should read that as an instruction to keep a copy of anything you genuinely cannot lose.
A week of daily restore points
Taken at the hypervisor, so nothing has to be installed inside the machine and a compromised guest cannot talk the agent out of it. The oldest point rolls off as the newest lands, which is what “rolling” means and why the window is exactly seven days rather than seven of the days you liked.
20 % of the instance, per month
An instance with the backup add-on bills 120 % of the instance — the add-on is a fifth of the line, and the GST split is shown on the invoice as it is everywhere else. It is the same arithmetic as the rest of the product: a percentage of a number you already know.
It is not a recovery guarantee
The service agreement disclaims any guarantee of recovery from the backup add-on, and we would rather print that sentence on the product page than leave it for the day you need it. Treat it as a fast undo for your own mistakes, and keep an off-platform copy of anything irreplaceable.
If the requirement is a real backup product with its own retention, versioning and immutability, that is a different job and we sell a different thing for it. See Comsky Backup →
S3-compatible buckets, over the API today.
Buckets, per-bucket credentials and short-lived signed URLs are reachable through the Comsky API now. There is no console screen for them yet, so this is written as what it is rather than dressed up as a shipped product surface.
Per-GB rates for object storage are published at launch. Ask us before you design around it. Browse the API reference →
The rest of it, without the paragraphs.
Root disk
- Local NVMe on the same hypervisor
- 10–16,384 GB per instance
- metered per GB per month, GST inside
- 1–50 GB of SSD per GB of memory
- Grown by resizing the instance
- Never leaves ZONE01
Volumes
- Created independently of an instance
- Attached, grown, detached, deleted
- Priced with a quote before creation
- Moved between instances
- Cut fresh from a snapshot
- Rate quoted rather than published
Snapshots
- Point-in-time copies on demand
- Volume and whole-instance snapshots
- Revert a volume onto a snapshot
- Create a new volume from one
- Taken before a risky change, in seconds
- Deleted when the change stuck
Backup add-on
- Daily policy, 7 rolling restore points
- Enforced at the hypervisor
- 20 % of the instance price per month
- Provisioned with the instance
- Opt-in, and removable
- Not a guarantee of recovery
Object storage
- S3-compatible buckets
- Per-bucket access key and secret
- Short-lived signed URLs per operation
- Any S3 client you already use
- Object data never crosses our API
- API today — no console screen yet
Data movement
- No egress metering
- No bandwidth allowance to exceed
- Snapshots and volumes stay in ZONE01
- Your own ISO uploaded to the library
- Restores land on a new instance
- Root on the machine holding the data
About storage specifically.
Is the instance SSD local or over the network?
Local NVMe on the same hypervisor the instance runs on. That is why the SSD is a line in the shape at metered per GB per month rather than a separately provisioned network volume you wait for.
What does the 7-day backup actually protect me from?
Your own mistakes, mostly — a bad migration, a deleted table, a change that looked fine on Friday. It gives you seven daily restore points held at the hypervisor. It is explicitly not a guaranteed recovery service, and the SLA says so, so keep an independent copy of anything irreplaceable.
How much is the backup?
20 % of the instance price per month, so an instance with it bills 120 % of the instance. The invoice shows the taxable value and the GST inside that line as it does everywhere else.
How is a snapshot different from a backup?
A snapshot is one copy you took, when you chose to, and it stays until you delete it. The backup is a policy that takes one every day and drops the eighth. Use the snapshot before a change you are nervous about; the backup is for the morning you did not see it coming.
Can I restore without touching the broken machine?
Yes, and it is the better move. Cut a new volume from the snapshot, deploy a clean instance, and attach it there. The instance you are worried about stays exactly as it is while you look at it.
Can I grow a disk without downtime?
An attached volume can be grown in place without detaching it or stopping the instance. Extending the filesystem inside the guest is yours to do — you have root, and we do not reach into your machine to do it for you.
What do I pay for data going out?
Nothing. Egress is not metered and there is no transfer allowance to exceed, including on the day you decide to take everything and go.
What does an extra volume cost?
We quote it rather than publish it, and the quote endpoint prices a specific volume before you create it, and the a per-GB monthly rate for the instance SSD is live today.
Start with Comsky Cloud.
Grow into the ecosystem.
One account opens every Comsky product — and one invoice covers them, however many you run.