I can see a timeout in our Python history, but the PLC and controller clocks disagree with it and the application reused its job label after restarting; how should I reconstruct only the supported ordering and leave the acknowledgement unassigned if the retained evidence cannot identify its attempt?