Terraform locking is not an AWS provider feature.
The backend acquires the lock before infrastructure changes begin.
In this lesson, I break down how CI/CD jobs request locks, how long locks remain active, and why stale locks become operational risks inside shared pipelines.
The lesson also explains why backend infrastructure must exist before terraform init, how IAM permissions affect lock ownership, and why force-unlock can create concurrent state mutation conditions.
The focus is mechanism first:
lock acquisition, lock duration, state protection, and failure boundaries.