Workspaces now wire themselves up

If you split your infrastructure across multiple workspaces — networking, IAM, data, app — you almost certainly use terraform_remote_state to pull outputs across boundaries. And you've almost certainly had to also tell InfraDots about those dependencies by hand so that "apply VPC → re-apply consumers" works.

That second step is gone.

What's new

Flip the new Auto-discover Workspace Interconnections toggle in your organization settings. From the next state upload onward, InfraDots:

1. Parses the uploaded state for terraform_remote_state data sources (backends remote and cloud).
2. Resolves the referenced workspaces inside the same organization.
3. Creates or updates a FULL_APPLY interconnection on the source workspace.

Any interconnections you previously created by hand are left untouched — auto-discovered rows live alongside manual ones and only the auto rows get re-synced on each state upload.

Why it matters

The old failure mode looked like: you add a new data "terraform_remote_state" block, you forget to wire the dependency in InfraDots, the upstream workspace gets applied, and downstream consumers happily read stale outputs until someone notices a drift report two days later.

With auto-discovery, the dependency graph follows the code. Your state is the source of truth.

Also in this release

While shipping this we fixed a long-standing bug where FULL_APPLY interconnections never triggered downstream applies — the guard was comparing a Job object to a status string and always tripping. If you were relying on FULL_APPLY rows and noticed they "didn't seem to do anything," that's why. They do now.

Turning it on

Settings → General → toggle Auto-discover Workspace Interconnections. Off by default, opt-in per organization, no migration required.