Technology
Terraform
Modular infrastructure code with a state strategy that survives multiple teams — and no console changes that nobody remembers making.
If it is not in code, it does not exist
Console drift is how environments become unreproducible. The eventual cost lands during an incident, when the disaster-recovery region turns out to be subtly different from production.
We describe everything in Terraform, keep modules small and versioned, and split state so a change to one team's infrastructure cannot lock or damage another's.
Practice
How we use it
Module design
Small, versioned, composable modules with sensible defaults, so teams consume rather than copy.
State strategy
State split along team and blast-radius boundaries, with locking and remote backends as standard.
Policy as code
Automated policy checks in the pipeline, catching a misconfiguration at review rather than in an audit.
Drift detection
Scheduled plan runs that surface manual changes before they cause a surprise during a deploy.
Judgement
When Terraform is the right call
And when it is not. A technology page that only lists strengths is a brochure.
Reach for it when
- Multi-environment, multi-account estates
- Anywhere reproducibility or audit evidence is required
- Teams needing a review gate on infrastructure change
Look elsewhere when
- Application-level configuration, which belongs closer to the application
- Genuinely single-environment prototypes with a short life
Sectors
Where we run it
Also relevant