Onkor shows a similar 45 minute gap in my flow recording on Saturday early am, around 3 am also. The same data viewed in Sleepyhead shows only ~ a 5 minute gap, and I know I had a bathroom break right there (after waking up with a really dry mouth). I think maybe there are a couple of flaky values in the flow data that Onkor can't handle, so it just leaves a gap in the flow until it finds something interpetable. I'll try to get the plots posted here from both Onkor and SleepyHead so you can see the difference...
Here's Onkor zoomed in on the gap

and here's SleepyHead, same time frame (but slightly different scaling)

Strangely, neither program seems to get it completely right! And the only break I remember is the one at about 3:22-3:25
(this is of course SleepyHead 0.8.4 that jedimark is still working on, so there may be graphing anomalys there... but I thought Onkor was pretty stable and reliable?)