I think you are onto it. If the gland failure happened once already under water, then no additional air could get into the housing, thus not raise the pressure (decrease the vacuum) enough to trigger the alarm.
But another interesting issue raised above by junior5 - once sealed, how is it possible for the pressure inside a rigid housing to significantly change at depth? That would seem to violate the laws of physics. Not counting temperature effects of course.
The only way the pressure can change is either a huge temperature change, not realistic, or something leaks into the housing to displace volume, and enough has to leak in to compress the air in the housing enough to make a big enough pressure change to trip the alarm. In other words a fair bit of water.