Setting up
New servers at Hetzner, OVHcloud, Scaleway, AWS or Azure, or on hardware you already own. Same hardened baseline every time, and it's documented.
Managed servers
Managed Linux servers, watched by AI agents
You keep the server (or we rent one in your name). We set it up properly, patch it, back it up and keep an eye on it. The watching is done by AI agents, because they don't mind reading logs at four in the morning. Changes are made by us, and written down.
What the agents actually do
Most outages we see are dull: a disk full of logs, a certificate nobody renewed, a process that hung after an update. The agents look at every server once a minute, work out the likely cause and prepare the fix from a runbook. A short list of small fixes (restart a crashed service, clear old logs) they may do on their own. Everything else waits until one of us has read it.
What's included
What we do
Updates
Security updates and kernel patches in an agreed window, plus emergency patching when something like regreSSHion comes out. Configuration lives in Ansible.
The software on top
Nginx, Apache or LiteSpeed, PHP, Node.js, Python, Docker, MySQL, MariaDB, PostgreSQL, Redis. Tuned for what the server actually does.
Watching it
Disk, memory, services, certificates and the pages your customers use, checked every minute. Real problems reach a person according to the plan you're on.
Backups
Encrypted, stored off the server, kept for as long as agreed. We restore one on a schedule to prove it works.
Moving in
From shared hosting, an old server or another provider. We plan the move so the downtime is minutes. Servers someone else built get audited first.
What you have at the end
- Servers set up to a written baseline
- A runbook for each one
- A log of every change
- A monthly report
What we don't do
- Hosting resale. You pay the provider directly
- Writing your application code
How it works
What happens when something goes wrong
The agents use the same runbooks we would, through an account that can only run commands on an approved list. You decide with us which small fixes they may do without asking.
- 1WatchDisk, memory, services, certificates and the pages customers use, checked every minute.
- 2Work out whyThe agent lines the alert up with recent changes and logs, and writes down the likely cause.
- 3Suggest a fixTaken from the runbook, together with the way back if it goes wrong.
- 4A person says yesOr no. Only the small fixes you agreed to beforehand skip this step.
- 5Apply itThe change goes in and the checks run again. If they fail, it’s rolled back.
- 6Write it downIn the change log, and in the report you get at the end of the month.
- web-02disk /var at 91%agent: looking
- web-02cause: journal not rotated (6.2 GB)
- web-02fix: vacuum journal to 1 GB, set limitsmall fix · agreed in advance
- web-02done · disk /var at 54%logged
- db-01replica lag 48 sagent: looking
- db-01cause: long-running report query
- db-01fix: send reports to replica-2needs a person
- db-01approved · appliedlogged
Plans
Three plans, chosen per server
Per server, per month. The hosting itself is extra: it's in your name, or we pass it on at cost.
Essential
Test servers and internal tools
€89 per server / month
- Agents watching
- Day and night
- Security updates
- Once a month
- Backups
- Daily, kept 14 days
- Restore test
- Every quarter
- When an engineer looks
- Next working day
- Report
- Monthly
Our usual recommendation
Professional
Websites and shops in production
€179 per server / month
- Agents watching
- Day and night
- Security updates
- Every week
- Backups
- Daily, kept 30 days, off-site
- Restore test
- Every month
- When an engineer looks
- Same working day
- Report
- Monthly
Critical
Systems that cost money when they're down
€279 per server / month
- Agents watching
- Day and night
- Security updates
- Every week, plus emergencies
- Backups
- Database hourly, full daily, off-site
- Restore test
- Every month, plus a failover test
- When an engineer looks
- Evenings too, ahead of the queue
- Report
- Monthly, with a call
Prices
One-off jobs and extra hours
Taking a server on
€290 per server
Audit, bringing it up to our baseline, documentation. Paid once.
Changes you ask for
€96.80 / hour
Work outside the plan, billed in half hours.
Prices are without VAT. Anything else is quoted, or charged at €96.80 an hour. All prices.
Software
What we use for this
- Ubuntu
- Debian
- AlmaLinux
- Red Hat Enterprise Linux
- Ansible
- Docker
- Nginx
- Apache
- LiteSpeed
- PostgreSQL
- MySQL
- Redis
- Prometheus
- Grafana
Questions
Things people ask
- Can the agents change things on their own?
- Only the small fixes you've agreed to beforehand, like restarting a crashed service or clearing old logs. Everything else waits for one of us. Both kinds end up in the change log.
- Do we keep root access?
- Yes. It's your server. We document every account and key we use, and you can take it back whenever you like.
- Which providers do you work with?
- Anything that gives you a normal Linux server. We'll suggest one if you haven't chosen, based on price, location and how much uptime you need.
- Someone else built our servers. Can you take over?
- Yes. We audit them first, fix what needs fixing, then they go on a plan like any other server.
Articles
If you want the detail
Managed servers
What we let AI agents do on a server, and what we don't
Agents are good at watching servers, linking up symptoms and writing the fix. The hard part is deciding what they may do without asking. The four groups we use, and the safeguards around them.
Managed servers
Locking down SSH on a server that faces the internet
regreSSHion and Terrapin were reminders that SSH needs looking after like anything else. The baseline we put on every server: patches first, keys only, no root login, a short list of who may log in, and an audit we can repeat.
Security & continuity
Your backup job says OK. That doesn't mean you can restore
A backup job that finishes without errors tells you something was copied. Whether you can get a working system back, and how long it takes, you only find out by restoring it. Here's how we do that, and what we write down each time.
Got something that needs looking after?
Tell us what you have and what's bothering you. You'll hear back within one working day, usually with a few questions before any price.