Skip to content

Tech stack

Boring tools, chosen on purpose.

We pick well understood tools, keep them in code and only add complexity when a real problem asks for it. Here is what we use and why.

How we choose

  1. 01

    Everything in code

    If it cannot be rebuilt from the repository, it does not exist yet.

  2. 02

    Managed before self hosted

    We only run what we must. Every server you skip is one less thing to patch at night.

  3. 03

    Watch it from day one

    Monitoring and alerts ship with the first deploy, not after the first outage.

  4. 04

    The smallest thing that works

    Simple systems fail in simple ways. Complexity has to earn its place.

What we reach for, and why.

Not a logo wall. Each row is something we use in production, with the reasoning behind the choice.

01

Cloud

  • AWS

Our default unless you are already committed elsewhere. The deepest catalogue of managed services, and most engineers you hire will already know it.

02

Containers and orchestration

  • Docker
  • Kubernetes

Docker for packaging, always. Kubernetes only once a workload genuinely needs orchestration, not because it looks good on a diagram.

See it in Kubernetes research platform
03

Infrastructure as code

  • Terraform
  • Ansible

Terraform provisions, Ansible configures what runs on top. State and playbooks live in your repository, not in someone else’s dashboard.

04

CI/CD

  • GitHub Actions
  • Jenkins

GitHub Actions for anything new, because it sits next to your code. Jenkins where a team already runs it well; we improve it rather than replace it.

See it in Budget Guardian
05

Edge and delivery

  • Cloudflare

DNS, TLS and caching at the edge. Fast pages everywhere and a smaller surface exposed to the internet.

See it in Budget Guardian
06

Observability

  • Prometheus
  • Grafana
  • Elastic Stack

Prometheus and Grafana by default. Elastic Stack joins once logs need to be searched at scale, not just charted.

07

Systems and scripting

  • Linux
  • Networking
  • Bash
  • Python

Linux and networking are the floor everything stands on. Bash for glue, Python the moment a script deserves tests.

See it in Automated content pipeline

Not on AWS?

The practice travels. The cloud is a detail.

Infrastructure as code, clean pipelines and good observability look the same on any provider. Tell us what you run and we will say plainly whether we are the right fit.

  • Google Cloud
  • Azure
  • Your current stack