Linux system administration and support: server management, hardening and automation

Linux system administration is the core of what we do: designing, configuring and keeping healthy the servers that run websites, applications and e-commerce. We don’t sell hosting as a sealed box - we work directly on your systems, document them and hand them back to you in a state you can actually understand. “We write the code and keep it running in production”: this page tells the second half of that sentence.

Need the first half instead - developing or evolving an application? Take a look at custom software development.

Overview

A well-managed Linux server doesn’t draw attention: it responds fast, updates without surprises, and when something breaks there’s a backup that actually restores. A badly managed one draws plenty of attention - usually at the worst possible moment.

The service covers the machine’s whole lifecycle: initial provisioning, hardening, planned updates, monitoring and intervention when something’s off. We work on bare-metal servers and virtual machines on any provider, either as one-off jobs - a problem to fix, a migration, an audit - or as ongoing management on a monthly retainer.

The method doesn’t change: standard configurations anyone can read, a single point of contact who genuinely knows your infrastructure, and reports that make sense even if you’re not a sysadmin. And what we configure stays yours: no proprietary panel, no artificial lock-in to us.

Typical problems we solve

• The server is slow and nobody knows why → we analyze load, I/O and configuration and remove the actual bottleneck, not the assumed one.
• The disk fills up and services stop → we set up log rotation, automatic cleanup and alerts that fire before things lock up.
• Updates have been frozen for years out of fear of breaking things → we schedule patches and distribution upgrades with a rollback plan ready to go.
• Backups exist but nobody has ever restored them → we make them encrypted, verified, with a restore procedure tested on a real machine.
• SSH exposed, permissions too broad, no firewall → we reduce the attack surface with hardening documented step by step.
• The previous provider vanished, taking the credentials with them → we run an audit, regain control of access and document everything.
• Every server is hand-configured and different from the rest → we bring the configuration into Ansible and make it reproducible.

What’s included

• Provisioning and installation: server setup from scratch - partitioning, users, services and the application stack (web server, PHP/Node.js, database, cache) - ready for production.
• Security hardening: key-only SSH access, firewalling, automatic security updates, fail2ban, correct permissions and an attack surface trimmed to what’s necessary.
• Updates and maintenance: regular system patching, planned distribution upgrades, log rotation and disk-space management.
• Automation: reproducible configurations with Ansible and idempotent scripts, so the environment can be rebuilt in minutes instead of days of manual work.
• Backups and recovery: encrypted backups tested periodically, with a restore plan we know how to run - not just configure.
• Monitoring and alerts: checks on resources, services and certificates, with notifications that arrive before your customers’ complaints do; on-call cover on a monthly retainer when downtime actually costs you.

Stack and technologies

We work daily with Debian and Ubuntu Server; on request we also manage AlmaLinux, Rocky Linux, Arch Linux and other distributions. On top runs the typical web production stack: Nginx or Apache, PHP-FPM, Node.js, MySQL/MariaDB or PostgreSQL, Redis for cache and sessions, Varnish where it earns its keep, Docker when containers genuinely simplify the deploy.

For automation we use Ansible; Terraform comes into play only where infrastructure complexity justifies it, never on principle. On the cloud side we work mainly with Google Cloud, alternatively AWS and DigitalOcean; on request also Azure, Hetzner and OVH. If you already have a stack, we adapt to it; if you’re starting from scratch, we pick the simplest solution that holds the load - and document it, so that a year from now anyone can understand why it’s built that way.

A real-world example

A web agency handed us about ten Ubuntu servers accumulated over years of projects: each configured by hand, different versions, updates lagging behind, and one night’s downtime caused by a disk saturated with logs. We started with an audit and an inventory of services and dependencies, then brought the configuration into Ansible: users, SSH hardening, firewalling, log rotation and automatic security updates, identical everywhere. The backups, which existed but had never been restored, are now encrypted and proven with a periodic restore on a test machine. Today a new server is ready in half an hour starting from the playbooks, alerts arrive before the clients do, and emergency work has become the exception: almost everything is planned, not scrambled.

Who it’s for

• SMEs and professional firms that want a reliable technical point of contact without setting up an in-house IT department.
• Web agencies and software houses that hand projects to clients and need solid servers behind the scenes.
• E-commerce and SaaS that can’t afford downtime in the moments that count.
• Anyone who’s inherited undocumented servers and wants to regain control before the next problem hits.

How we work

The first step is an audit: we take a snapshot of the systems’ state, flag the risks and propose a plan with clear priorities. Then we work in stages, without tearing down what already works, and leave every configuration documented: if you wanted to switch provider tomorrow, you could - and that’s exactly the guarantee that we’re doing good work today.

Standard coverage runs Monday to Friday, 10:00-18:00 , with priority triage: a P1 is picked up within one business hour (details are on the response times page). For pay-as-you-go work there are hour packages that never expire; quotes are always free, with a reply within one business day.