A shared Docker infrastructure hosting all of our projects

This infrastructure is the technical foundation every Kolixia project runs on, including this website and our own SaaS products. It reflects our expertise in Docker architecture, observability and production reliability, skills we also put to work for clients on their own platforms.

The context

<p>Kolixia builds and hosts several projects in parallel: websites, business applications, SaaS products like Info Bot or Mealy. Every new project needs a database, a cache, HTTP routing, and visibility into its production health.</p>

The problem

<p>Hosting each project in isolation would have multiplied infrastructure costs and maintenance overhead, with no resource sharing and no centralized view of the health of every service running in production on the machine, making every incident slower to diagnose.</p>

Our solution

<p>We built a Docker infrastructure organized around 3 complementary pillars, all connected to the same external Docker network. A data layer centralizes PostgreSQL, Redis, Neo4j (exposed in development for technical dependency mapping), pgAdmin and Flower for every integrated project. A monitoring layer, built with Grafana, Prometheus, Alertmanager, Loki, Tempo, Grafana Alloy, Node Exporter and cAdvisor, watches each project's activity as well as the host machine. A core layer, built on Traefik, routes every newly integrated project and every service across the whole infrastructure, with automatic TLS certificates.</p>

The benefit

<p>Full sharing of critical resources (a single Postgres instance, a single Redis, a single reverse proxy for every project), onboarding a new project down to a few Traefik routing rules, and centralized visibility into the health of every service running in production, with Grafana dashboards checked daily.</p>