Erik is the founder of Cloud Posse and creator of Atmos. With over a decade of experience helping teams adopt Terraform at scale, he is passionate about open-source infrastructure tooling and developer experience.
Atmos Pro now tracks a foreign-key relationship between every workflow run and the infrastructure instance it operates on. When a plan or apply executes, the run is permanently linked to its stack-component pair — giving you a complete deployment history per instance, not just a flat list of workflow executions.
This means you can answer questions like "what was the last successful apply for this component?" and "which PR introduced this instance?" directly from the instance view.
Tracked runs unlock a new dimension of drift detection. Previously, drift checks always compared your infrastructure against HEAD of the default branch. Now Atmos Pro supports three configurable modes per repository:
Code drift compares live infrastructure against HEAD of main — the classic check that tells you whether your deployed state matches what your code expects.
Deployment drift compares live infrastructure against the exact commit SHA that was last successfully applied. This catches manual changes made after a deployment, even when no new code has been pushed.
Both runs code and deployment drift independently, with separate status tracking for each. You see at a glance whether drift is caused by new code that hasn't been applied yet, by manual infrastructure changes, or by a combination of both.
Instances now progress through lifecycle states — from proposed (created when a PR introduces a new component) through active (confirmed on the default branch) to orphaned (removed from code but still tracked) and archived. Proposed instances are automatically garbage-collected after a configurable TTL, so your instance list stays clean without manual intervention.
