Previously, the Orchestration service did not take stack action into account when checking the stack state. As a result, deleting a stack in a FAILED state would first cause the stack to go into a DELETE_FAILED status. It took a second delete attempt to actually delete the stack.
With this update, the stack action is now taken into account when checking the stack state. Specifically, the current action is now compared with whatever action leads to a reported stack state. As such, stacks in a FAILED state will be correctly deleted on the first attempt. (BZ#1256564)