How does drift detection behave when "Ignore stack state" is enabled?

By default, Spacelift only schedules drift detection when a stack is in a finished state to avoid conflicts with other runs. If you see a message that drift detection wasn't scheduled because the stack isn't in a finished state, this is the expected behavior.

Enabling "Ignore stack state"

If you enable the "Ignore stack state" option in your stack's Drift Detection settings, Spacelift will run drift detection even when the stack isn't finished. When this setting is enabled, drift detection will start immediately as long as there's an available worker in the assigned pool.

What happens when drift is detected?

When drift is detected and reconciliation is disabled, Spacelift will:

• Create a proposed run showing the detected changes

• Not apply the changes automatically

• Not continue adding new proposed runs for the same drift while an existing proposed run is unresolved

Drift detection will resume according to its schedule only after the proposed run is approved, discarded, or replaced by another tracked run, and the stack returns to a finished state.

For more detailed information, see the Drift detection in practice section in our documentation.