How long until the server is running again?
Not: until the files are back in a folder.
- Bootable recovery ISO with a graphical wizard
- Sign-in with a recovery code – no password from the dead server required
Warden Backup
So the day after an incident is just another working day
Enterprise backup with end-to-end encryption, ransomware detection and three separate copies. You're not buying storage – you're buying a tested route back into operation. And we show you that it works before you need it.
Encryption
The key stays with you
Copies
3 copies, separate servers
Recovery
Bootable, onto new hardware
Billing is based on the original size of your data – a number you can measure and compare yourself. All prices incl. 19% VAT.
Three questions for your current backup
The price of a backup is never compared with the price of hard drives, but with the day it is needed.
Who can read your data if our data center is compromised?
No one without your key – not even us.
- Encryption happens on your machine, before a single byte leaves your premises
- There is no button that opens it – that's the architecture, not a promise
How do you know last week's backup is still good?
Silent data corruption is why a backup that sat on the shelf for years sometimes no longer works.
- Every piece of data is re-verified regularly
- A damaged piece is automatically replaced from a healthy copy
- Verification jobs: run a test restore without overwriting anything
Enterprise-grade backup
Built for companies that need to be back at work right after an incident – not eventually.
Recovery is the product
Anyone can back up. The difference shows when the server is dead.
- Single files in seconds via the web interface
- Back onto different hardware – the target is checked beforehand, not halfway through
- Complete system onto bare metal, verified byte by byte
- Full-disk backup sector by sector – including boot sector and partition table
End-to-end encrypted
The key stays with you; we read nothing.
- Modern, well-established encryption
- Recovery key for the case that server and agent are gone
- The restore leaves no key behind on the target system
Three separate copies
Not three disks in the same server.
- Every piece of data on three separate storage servers – currently within one region; further sites are in preparation
- If one copy fails, Warden fetches the piece from the next – without any action on your part
- Storage servers are replaced during operation, without maintenance windows
Ransomware detection based on the data
Not just at the network – but on the backups themselves.
- Deduplication rate: an encryption run rewrites every file, so nothing is recognized anymore
- Compression rate: encrypted data can no longer be compressed
- On suspicion, older versions are protected from retention cleanup
- WORM lock at two levels: governance (default – early deletion only by us, and only as a logged override) or compliance (no early deletion at all – not even by us)
System monitoring, optional
The agent is running anyway – on request it does more than backups. Bookable per machine.
- New root accounts, changed SSH keys, newly opened ports
- Unknown and short-lived processes, modified scripts in the web root
- Regular outbound connections at a steady beat
- On a critical finding: automatic security snapshot with a stated reason
Proof at the push of a button
Built for the auditor, not for the brochure.
- Compliance report as PDF: GDPR Art. 30/32, ISO 27001 A.8.13
- The report explicitly lists its gaps – failed backups are not concealed
- Audit log of all security-relevant actions
What can be backed up
A recipe knows how to cleanly pause, back up and release an application – you don't write a script.
Files & foldersAny number of paths, exclusion rules, locked files
Entire disksSector by sector via snapshot – LVM and VSS
DatabasesMariaDB / MySQL, PostgreSQL as a consistent dump
Ready-made recipesNextcloud, Plesk, MongoDB, Redis, Docker volumes
VirtualizationVMware (vSphere) and Proxmox via their APIs
Cloud servicesMicrosoft 365 and Google Workspace
PlatformsLinux and Windows, each on amd64 and arm64
Custom scriptsPre- and post-run hooks per plan
TransferDeduplication and compression always on – short backup windows, low bandwidth usage
Shared responsibility with Microsoft 365 and Google Workspace: the providers are responsible
for the availability of the service – backing up your content and being able to restore it is your
responsibility under their own shared responsibility model.
Pricing model
Tiered by storage volume – like a tax table: the first terabytes at the base rate, anything above at the next tier.
| Tier | Volume | € / GB | € / TB |
|---|---|---|---|
| 1 | 0 – 2 TB | €0.0250 | €25.59 |
| 2 | 2 – 10 TB | €0.0232 | €23.76 |
| 3 | 10 – 25 TB | €0.0196 | €20.11 |
| 4 | 25 – 100 TB | €0.0155 | €15.84 |
| 5 | from 100 TB | €0.0119 | €12.19 |
Base fee per protected machine€5.94 / month
HA plan, in addition to the base fee€9.51 / month per plan
System monitoring per monitored machineon request
Billing is based on the original size of your data – a number you can measure and compare yourself. All prices incl. 19% VAT. Volumes use binary units: 1 TB = 1,024 GB, the same way Windows and common backup tools display them. An HA plan keeps three copies on separate storage servers; storage is billed per copy, so three times. Distribution across several sites is in preparation.
Billed on original size
A single number you can measure yourself.
- Full backup plus the changed data of each subsequent version
- Your price depends on your data volume, not on our storage internals
- So it doesn't fluctuate if your data suddenly compresses differently – for example after a format change in your database
What else the invoice shows
Every line can be recalculated.
- High availability is itemized separately – three copies are three copies
- Backups in the trash are not charged
- Retention counts in chains: “keep ten” means ten restorable versions
Frequently asked questions
The most convincing argument is a demo: one machine, one test run, one restore from the ISO.
Request a demo
Because that comparison puts storage next to a backup. Raw storage doesn't back up, doesn't verify,
doesn't detect attacks, doesn't boot a server and doesn't produce a report for the auditor.
What you get there is a shelf – what you get here is what belongs on the shelf,
and someone who regularly checks that it's still there.
Two questions decide that. Is the NAS in the same building? Then it won't survive fire,
water damage or burglary. And: if your server gets encrypted tonight – the NAS is attached
to it as a network drive. With us, backups are stored three times on separate storage servers, not deletable within the
defined retention period, and encryption of the source data is detected on the next run.
The real danger isn't the server outage, but the backup copying the encrypted files for four weeks
while retention then discards the last healthy versions.
Warden measures two values it determines for every backup anyway: if the
deduplication rate collapses, every file was rewritten. If the
compression rate collapses, the source is encrypted. On suspicion, the
older versions are protected from retention cleanup and an alert is raised.
Only in part. Microsoft is responsible for the availability of the service; backing up your
content and being able to restore it is your responsibility under the shared responsibility model.
A deleted mailbox is gone once the retention period expires,
and a ransomware infection synchronized via OneDrive is synchronized – and broken.
The same applies to Google Workspace.
That's exactly what the recovery key is for. With it, a freshly installed
agent can access the old backups. Without it, nobody could get to them anymore – not even us.
That's not a shortcoming but the promise: we cannot open your data.
Then don't take them. High availability is a separate line item and itemized separately –
you see exactly what it costs and decide for yourself.
We'll demonstrate it before you sign
One machine, one test run, one restore from the ISO. That's the argument no data sheet can replace.