If the events tend to be too long, the calibration script (and sometimes TS generation) don't work. Working to determine what the limit is. 100,000 minutes is definitely too long.
Fixing a unit issue (with peak flows, dividing volume by peak flow CFS and assuming the result to be in minutes) seems to keep the durations sane, so it's not a critical issue.
If the events tend to be too long, the calibration script (and sometimes TS generation) don't work. Working to determine what the limit is. 100,000 minutes is definitely too long.