Thanks for looking into this for me.
This was what the setpoints were doing last night when the problem was occuring:
The longer periods were when the DHW run was actually happening.
I’ll see if I can catch it doing it later when it repeats and telnet the output.