We back up everything on our servers automatically, around the clock. You don't have to set anything up or switch anything on – it's part of your hosting.
That said, our backups are a safety net, not a replacement for your own. Please read the section "Why you should still make your own backups" below.
What we back up, and how often
Every hour, we take a snapshot of all data on our servers using ZFS or Ceph. A snapshot is a complete, frozen-in-time copy of the file system – it captures your website files, databases and mailboxes exactly as they were at that moment.
These snapshots are transferred to separate backup servers, so your data doesn't depend on the machine it normally runs on. In addition, a daily backup is stored in an external data centre.
How long we keep backups
We keep 90 days of history, in decreasing detail the further back you go:
Hourly snapshots for the last 2 days
Daily snapshots for the last 30 days
Weekly snapshots for the rest of the 90 days
In practice: you can go back to almost any hour of the past two days, any day of the past month, and any week of the past three months. Once a backup passes the 90-day mark, it is deleted automatically.
How to restore something
You cannot browse or download the snapshots yourself – they live on our backup infrastructure. Just write to support@dotplex.com and tell us:
what you need back (a whole server, a single website, a database, a folder, one file, a mailbox …), and
when it was still intact (a date, or a date and time – the more precise, the better).
We can restore an entire virtual server, or pull out individual files and data and hand them back to you. Whenever possible we'll place the recovered data next to your current data rather than overwriting it, so you can check it before it replaces anything.
Tip: Tell us as soon as you notice a problem. Recent data is available at hourly resolution, older data only weekly – so the sooner you ask, the closer we can get to the exact state you want back.
What it costs
Nothing. Restoring data is included in your hosting, and we won't send you an invoice because something went wrong.
We do, however, reserve the right to charge if restores are requested regularly. The reason is simply that a restore is manual work on our side, and it can take a while. If you can't tell us exactly when the data was last intact, we have to open several snapshots one after another and check each one to find the last good version. Pulling out individual folders or a single mailbox – rather than rolling back a whole server – means unpacking a snapshot and picking out the right pieces by hand.
For the occasional accident that's absolutely fine and included. It only becomes an issue if restores turn into a routine – at which point it's usually cheaper and faster for everyone if you set up your own backups, which you can restore yourself in minutes.
Why you should still make your own backups
Our backups protect against hardware failures, server problems and disasters on our side. They are not designed to be your day-to-day undo button, and they can't cover everything – for example, if something goes wrong and nobody notices for four months, the relevant backup is already gone.
Under our Terms and Conditions (section 7.10), keeping your own current, complete and working backups of everything you can access is your responsibility. Specifically:
Make a backup before every change you make yourself – updates, plugin installs, template changes, bulk edits.
Make a backup before announced maintenance by dotplex or our contractors.
Keep your backups somewhere other than the dotplex servers – storing your own backup copies on our servers is not permitted (and wouldn't protect you anyway).
The big advantage of your own backups: you can restore them yourself, immediately, without waiting for anyone. Most content management systems, shops and control panels have a backup function built in or available as a plugin – if you're unsure which one fits your setup, ask us and we'll point you in the right direction.
And if the hardware fails?
All our servers, routers and switches are redundant. If hardware fails, we start your website on a hot-standby server so that downtime stays as short as possible – separately from, and in addition to, the backups described above.
The full details are in sections 6.2.6 and 7.10 of our Terms and Conditions (AGB).