It always starts simple: a few cloud resources declared in Terraform, a small pipeline that authenticates to the cloud and does "cd $TF_PATH; terraform init; terraform apply".

But in 6 months, you find yourself managing 20 separate environments in Terraform, buried in a sprawl of subfolders, for teams that create new projects faster than you can say "terraform apply". Personally, I don't like having to maintain a hard-coded list of any kind, and even less a list of terraform environments owned by other people.

In this talk, I will show you how to create a universal, scalable and reusable Terraform pipeline with a little bit of magic:

- Auto-detect which environments have changed in a pull request (without having any lists!)
- Execute jobs in parallel thanks to matrix strategies
- Scan, validate, and detect security flaws in your code
- Require manual approval for production deployments

You will leave with all the knowledge you need for a pipeline that scales from 3 to 30+ environments without even touching the code!