What is OpenTofu, and does it affect us?

Terraform’s licence changed in 2023 and OpenTofu emerged as an open fork, now under the Linux Foundation, with HashiCorp itself since becoming part of IBM. For most users the two remain compatible. It is worth a decision rather than a default, though.

Do we need Terraform if we use one cloud?

The multi-cloud argument is the least important one. The value is that your infrastructure is written down, reviewable, and repeatable, and that applies just as much on a single provider. Portability is a side effect rather than the reason.

Who should be allowed to run apply?

Ideally nobody, directly. Changes should run through your pipeline after review, with people holding the ability to approve rather than to execute. That way every change to your infrastructure has an author, a reviewer, and a record.

How long does adoption take?

A first environment usually takes two to four weeks including the import work. Bringing a full estate across takes longer and is best done in stages, since a partial adoption that works beats a complete one that stalls halfway through.

Terraform
Development Services

Stateside Digitals provides Terraform development services, putting your infrastructure into code so environments can be reviewed, repeated, and rebuilt without depending on anyone remembering how it was set up.

Terraform describes your infrastructure in files rather than in a web console. Servers, networks, databases, permissions, and DNS records all become code held in version control alongside your application. Applying it builds what the files describe, and applying it again changes only whatever differs from what is already running.

For a business, the shift is that infrastructure stops being something only one person understands. Changes get reviewed like any other code, the current setup is readable by anybody who needs to see it, and the state of your cloud account no longer depends on somebody recalling what they clicked in a console eighteen months ago.

Infrastructure You Can Read

The Diagram Is The Code

What exists and why is written down and versioned, so nobody has to click through a console to work out how something was assembled.

Changes Reviewed First

See It Before It Happens

Every change produces a plan showing exactly what will be created, altered, or destroyed, which gets read before anything is applied.

Environments That Match

Staging Like Production

Environments are built from the same definitions, so a test that passes in staging is a meaningful signal rather than a rough one.

Environment Provisioning

Standing up new environments for a team, a customer, or a test run in minutes, identical each time and removable when finished.

Multi-Account Setups

Separate cloud accounts per environment or team, with networking, permissions, and guardrails applied consistently rather than configured by hand once for each new one.

Compliance Baselines

Encryption, logging, tagging, and access rules enforced in code, so an auditor sees policy applied rather than policy described.

Here is the question worth asking about your current setup. If your cloud account disappeared tonight, through a mistake, a compromise, or a billing dispute, how long would recreating everything exactly as it was actually take? For most businesses the honest answer is weeks, and it involves one or two people rebuilding it from memory while everyone else waits.

Infrastructure as code turns that answer into hours. Not because Terraform is clever, but because the description exists somewhere a machine can act on rather than in somebody’s head. The same property makes ordinary work safer: a new environment becomes a copy rather than a careful reconstruction, and staging genuinely resembles production instead of approximately resembling it.

We write modular Terraform with remote state and locking rather than files on a laptop, provisioning across AWS, Azure, Google Cloud, and Kubernetes. Plans run in your pipeline so changes are reviewed before they apply, with policy checks catching anything that breaches your own rules. Existing infrastructure gets imported rather than rebuilt, and drift is reported as it appears.

Terraform
OpenTofu
Remote State
State Locking
Modules
AWS
Azure
Google Cloud
Kubernetes
Policy Checks
Drift Detection
Import

State Stored Remotely

Plans Reviewed, Not Skipped

Modules Over Copy-Paste

Drift Detected Early

State lives somewhere shared and locked, so two people applying changes cannot quietly overwrite each other’s work.

The plan is read before every apply, because it is the only place that says out loud what is about to be destroyed.

Repeated patterns become modules, so a fix lands once rather than in the six places somebody duplicated it last year.

We check for changes made outside Terraform on a schedule, since drift found at apply time is drift found at the worst moment.

State Stored Remotely

Plans Reviewed, Not Skipped

Modules Over Copy-Paste

Drift Detected Early

Yes, and that is how most adoption goes. Terraform can take control of resources already running without recreating them, so nothing goes offline. We usually bring one environment across first, confirm it behaves, then work through the rest steadily.

Terraform
Development Services