This article reviews advances in detection and diagnosis of plant-wide control system disturbances in chemical processes and discusses new directions that look promising for the future. Causes of plant-wide disturbances include non-linear limit cycles in control loops, controller interactions and tuning problems. The diagnosis of non-linearity, especially when due to valve stiction, has been an active area. Detection of controller interactions and disturbances due to plant structure remain open issues, however, and will need new approaches. For the future, the linkage of data-driven analysis with a qualitative model of the process is an exciting prospect. Finally, the paper offers some brief comments about emerging applications.