A method is described for monitoring a plurality of process instances, such as business process instance. The method comprises: providing a model object reflecting a workflow common to the process instances, the model comprising a plurality of nodes with at least some of the nodes being capable of assuming at least active and blocked states; receiving events and changing the states of the nodes in corresponding model object instances in accordance with the events received so as to reflect the status of the process instances, repeatedly determining an “at risk” state indicator for the model instances, the “at risk” state indicator when set for a model instance indicating that the corresponding workflow instance is in a state in which it cannot progress to a normal completion without encountering a node in the blocked state.