Servers that don't crash at three on Sunday.
Configuration, migration and maintenance of Linux servers for serious sites and applications. Plesk when it adds value, plain Linux when it doesn't. No drama, with logs actually read and backups that actually restore.
Most server problems I see aren't from lack of horsepower, they're from careless configuration accumulated over the years: PHP-FPM with default values, MySQL with no tuning, ModSecurity too aggressive blocking legitimate requests, Imunify360 eating RAM, no proactive monitoring. When I step into a struggling server, 80% of the work is reading logs and understanding actual state, not installing new software.
I work mainly with two contexts. First, Plesk on CentOS or Ubuntu, which is what most of my clients use. I know its limits, its configuration quirks and its weak points across providers. Second, plain Linux when you don't need Plesk: nginx/Apache configuration, PHP-FPM per pool, MariaDB, supervisor for long-running processes, fail2ban, basic monitoring with Netdata or similar. Migrations between providers and between operating systems (CentOS 7 → Ubuntu 24, that kind of generational jump) I've done several times.
Something I learned the hard way: backups only count if you can restore them. Every server I maintain has a backup plan with periodic restore testing. Without that, you don't have backups, you have files you hope will work the day you need them.
How we work
-
01
Initial diagnosis
Server access, log reading, configuration review, identification of real bottlenecks. Before touching anything, understand what's happening.
-
02
Action plan with priorities
Concrete document: what to fix first (critical), what to optimise next (performance), what to refactor mid-term (architecture). Not everything is equally urgent.
-
03
Step-by-step execution with rollback
Each major change with prior snapshot, agreed maintenance window and documented rollback. Nothing is tested directly in production.
-
04
Monitoring and maintenance plan
Alert configuration (load, RAM, disk, downed services), periodic log and backup review, and optional maintenance plan so it doesn't degrade again.
What's included
- Initial server audit (no cost for recurring clients)
- Plesk and associated services configuration or reconfiguration
- Migration between providers or operating systems
- Tuning of PHP-FPM, MySQL/MariaDB, nginx based on real load
- Basic hardening (fail2ban, firewall, SSH hardening)
- Backup plan with restore testing
- Documented disaster recovery
- Optional monthly maintenance plan
When we're not the right fit
- You need Kubernetes or massive microservices Not my territory. If your real load needs orchestration of hundreds of pods, I'll point you to an SRE with that experience.
- "I want the cheapest server possible" Optimising hosting cents to the max usually gets expensive long-term. If that's your only criterion, better a basic managed VPS.
- You're after compliance certifications (ISO, SOC2) I can help with the technical side, but formal auditing needs a specialised consultancy that I'm not.
Stack & tools
- Linux (Ubuntu 22/24, CentOS 7/Rocky 9 when there's legacy)
- Plesk Obsidian (latest)
- nginx, Apache, PHP-FPM with tuned configuration
- MariaDB 10/11, MySQL 8, Redis for object cache
- fail2ban, ModSecurity, iptables/nftables firewall
- Backups with rsync/restic, off-site storage on a different system
- Monitoring with Netdata, alerts via email/Slack
If your server gives you problems, or you're thinking of migrating and need someone who handles it without you having to notice, write to me with what you have and we'll take a look.
Let's talk →