When there haven't been any earthquakes for a while (I think about 7 hours, but I sure of how to verify this), there are processes that clean up the event files that seismon generates and leave the directory empty. The way the epics code is written, this means the clock that we use for the diag_main test and "seismon dead" light on the compact seismon screen stops updating, which makes that diagnostic confusing.  I just noticed that last night, and I'm not sure what to do about it, but it's one of the few tests I can think of that are visible outside of the machine.  The only way to be sure at this point is check that the list mostly agrees with the USGS page (seismon spams some events, occasionally misses one) and watch to see if the diag_main message clears when a new eq shows up.