Can you also show a "power output" graph (or something like that, which show PWD duty cycle output to the hot end)?
Have you noted any relation between temperature raise and fall and the printing position?
Looking at graph I can only suggest the heating is suddenly stops and then the same sudden resumes.
It looks like you have a loose wire connection to the heater. When head is starting to move, it makes disturbance for the wiring, so heater occasionally stops to work.
There are usual two weak points: first it is where cable is connected to the board. Check is connector heated? Are any noises/sparks from there?
Another part: it is just at the point where wires going into the hot end. If wires are not secured to the printing head, then during printing those place experiences a lot of bending, which makes it weaker over the time.