When defining a median e.g. like Et_CO2 = median(Raw_etCO2[-1min, 0min]) the median is not computed correctly. See attached service Journal (Service-median01.zip) for etCO2 and MAC When defining a median like Et_CO2 = median(Raw_etCO2[now-1min, now]) everything is fine.
In this bug, the time database for newly set (but unchanged) input values are updated, so containing multiple entries with the same value. But the rules using this history haven't been updated, because the symptom itself has not changed, only its history.