Robot Studios IT Consultancy BV

Sheet 02 ยท Process line

Everything here is work I already do for my own software

The consultancy isn't a separate skill set. It's the same servers, the same framework and the same pager, pointed at your stack instead of mine.

Stations 01 to 06

What I take on

01

Linux servers

Provisioning, hardening and maintaining Ubuntu and Debian boxes. Web servers, PHP-FPM pools, MySQL, Redis, queues, backups, log rotation, ... The unglamorous half is the half that decides whether a server survives its third year.

02

PHP and Laravel

Building Laravel applications, and dragging old ones forward. If you're stuck three major versions behind because nobody wants to touch it, that's a job I've done more than once. Tests, static analysis and CI come with it, they aren't a line item.

03

DNS, TLS and mail

Zone design, migrations between providers, DNSSEC, certificate automation, and the SPF, DKIM and DMARC records that decide whether your mail lands in an inbox or a spam folder. I ran a DNS monitoring service for five years, so this is home turf.

04

Monitoring that stays quiet

Uptime, certificate expiry, broken links, cron jobs that stopped running without telling anyone. The hard part isn't collecting the data, it's deciding what deserves to wake someone up. An alert everyone ignores is worse than no alert.

05

Performance and capacity

Finding out why a page takes four seconds. Query profiling, caching layers, Varnish, HTTP/2 and HTTP/3, CDN configuration, and load testing before a launch rather than during it.

06

Security and incident response

Reviewing an app and its servers for the boring failure modes that actually get sites defaced: outdated packages, a database bound to 0.0.0.0, missing rate limits, a forgotten staging box. And when it already went wrong, working out what happened and closing it.

Operating procedure

How this usually goes

  1. Step 1

    A call and a look at what you have

    If it turns out somebody else is a better fit, you hear that on the call rather than three invoices later. I'd rather lose the work than take a job I'm the wrong person for.

  2. Step 2

    By the day, or fixed scope

    Billed in euro, 30 day terms. I work remotely from Belgium and I'm happy to be on site around Antwerp when a project genuinely needs it, which is less often than people expect.

  3. Step 3

    Everything gets written down

    Server changes land as committed configuration, not as a shell session nobody recorded, and the repository is yours when we're done. Nothing I build should need me to keep it alive.

Bill of materials

The stack I know best

It's a boring list on purpose. I reach for something exotic when the boring option genuinely runs out, and not a moment before.

Languages and frameworks
PHP 8, Laravel, Livewire, Vue, Tailwind CSS, Go, and a lot of Bash
Servers
Ubuntu LTS and Debian, nginx, Caddy, Apache, Varnish, HAProxy, PHP-FPM
Data
MySQL and MariaDB, PostgreSQL, Redis, ClickHouse, Elasticsearch
Operations
Ansible, Puppet, systemd, GitHub Actions, Zabbix, Prometheus, Grafana
Networking
BIND and PowerDNS, Cloudflare, Let's Encrypt, IPv6, firewalls

Something on that list giving you trouble?

Mail mattias@robotstudios.be with what you're running and what's going wrong. You'll get a straight answer on whether I'm the right person for it.