It seems there are a few minor issues with convergence tracking in the workbench. First, convergence tracking is triggered before a single run has been executed (which is superfluous). Second, convergence is not tracked one last time after the optimization has reached the stopping condition.