Well, of course there should be some feedback from a sensor in the heated mass (water) to control the heating source.
Back in the day I used a Ge probe, some op amps, and a timer to generate PWM that controlled a triac, or something like that. It was a long, long time ago, so I probably...