Hazardous events like stationary vehicles on the carriageway, being in most cases unforeseeable and not always easy to detect, pose serious challenges to automated vehicles (AVs). When such events occur, AVs have to determine within limited time and space if permanence in their Operational Design Domain (ODD) will be guaranteed or not, and how to react to ensure passengers' safety and comfort. To cope with such events more effectively and efficiently, in this paper we present a software architecture and logic for Connected AVs(CAVs) that takes into account hazard notification and road signage information from available standard V2X messages to manage ODD-related decisions and reactions in an anticipated way. Differently from earlier works, focusing more on automated compliance to traffic management suggestions by the connected road infrastructure, the presented solution emphasises the active role of the CAV logic in taking suitable decisions based on individual and local situations. We introduce a manoeuvre planner implementing distinct state machines to react to different types of received V2X information. In the resulting procedures, where the driver can be also involved, step goals for a motion planner and path controller are generated. By means of simulations, we demonstrate the benefits of the presented CAV solution against a baseline AV model only relying on on-board sensors. To prove its real-world feasibility, we also report the results of integrating the proposed logic into a CAV prototype and running real-world test-track experiments.