The Controller Area Network (CAN) is a communication bus for message transaction in small-scale distributed environments. Continuity of service and bounded and known message delivery latency are requirements of a number of applications, which are not perfectly fulfilled by existing networks, CAN included. One key issue with this regard, is that networks are subject to failures, namely partitions. However, most of non-critical applications can live with temporary glitches in network operation, provided these temporary partitions are time-bounded. We call these periods of inaccessibility. Should one call for hard real-time behaviour, a worst-case figure for these non-negligible periods should be derived and added to the worst-case transmission delay expected in the absence of faults. This paper does an exhaustive study of CAN inaccessibility characteristics, presenting figures for intervals in CAN operation when the network does not provide service, although not being failed.