The issue is that both scalers (or the analzyer) report a single scaler read of 60.439... x10^6. This throws off the auto-scale in the histogram. If I place a cut on either scaler read to trim this event then the data look good.
I don't know why this is happening now, nor do I understand how a 32 bit scaler can report a values >60x10^6. I suspect a change to the analyzer (Ole?)
It looks OK, so I updated the online config file with the cut so shift worker don't get confused, but it would be nice to understand this a little better. (Run 1595 shows this behaviour, for example.)
A copy of this log entry has been emailed to: moffit, ole