We run and monitor Linux servers on Ubuntu, Debian and AlmaLinux. Patches go in during agreed windows, and a restore test shows whether the backups work. The response time for incidents is in the contract.

Request a Legacy Check See service levels

# Maintenance window, Tuesday 06:00
web:~$ sudo apt update 2>/dev/null | tail -1
4 packages can be upgraded. Run 'apt list --upgradable' to see them.
web:~$ sudo apt full-upgrade -y -qq > /dev/null
web:~$ sudo needrestart -r a
Restarting services...
 systemctl restart nginx.service php8.4-fpm.service
web:~$ systemctl is-active nginx php8.4-fpm
active
active
web:~$ sudo needrestart -b | grep KSTA
NEEDRESTART-KSTA: 1
# kernel is current, no reboot needed
web:~$ curl -sI https://example.com | head -n1
HTTP/2 200
# done, goes into the report
web:~$

What we run

Ubuntu and Debian, plus AlmaLinux as the CentOS successor
Containers and Kubernetes with Docker, Coolify and Rancher with RKE2
Reverse proxies and caches with nginx, Traefik and Varnish
TLS with certificates and their renewal
Load balancers at Hetzner and Cloudflare in front of the website
Firewalls on the host and in the cloud
DNS records and zones, plus VPN access to the host
Backups, and for us a backup only counts as verified once we have restored from it
Monitoring for availability, logs and metrics, cron jobs and imports

If you want us to look after the application on the server as well, that is Lobster Legacy Care. Then one team takes care of code and server.

Patching on a fixed cycle

We keep an inventory, watch end of support dates and new CVEs and install patches in the maintenance window. Every round goes into the report. What we do in each round is under How we work.

Your service level sets the cadence. Basic has one maintenance window per month. On Business we patch every 14 days and apply critical patches within 72 hours. If a vulnerability is critical, we schedule it out of turn.

Move the mouse over a step or tap it. The step before and the one after stay highlighted.

Every patch round starts with the inventory and ends with the record in the report.

Events and alerting

We monitor with Uptime Kuma and also with UptimeRobot as a second, independent monitor. Alerts go to our team channel in Slack. For forms we test that a submission really arrives. A page that loads is not enough for us.

Cron jobs and imports report to the monitoring on every run. If that report does not come, it becomes an incident. A stalled import shows up before anyone misses the data. How we classify events and what happens after an alert is under How we work.

Upgrades and end of support

Ubuntu 22.04 leaves standard support in May 2027. For AlmaLinux 9 only full support ends on 31 May 2027. Security updates continue until 2032. These dates are in our EOL calendar, and we plan the move in good time.

A release upgrade is a normal change for us, with a risk assessment and a way back. The new server goes up next to the old one, we move the services across, and the old one stays in place until the new one runs. More under Modernisation.

What we do not do

We do not look after Windows Server, firewall appliances, office LAN and wireless networks or workplace IT. Your current IT provider stays responsible for those.