Yeah, I reached that conclusion myself as well. I think it makes sense to have a "Probe" device, with an Internal Temp child device and an Ambient Temp child device. The child devices will each be temperature sensors. The Probe parent device can then have data representing the cook as a whole.
As far as the time remaining aspect, I don't know that the Time Variable Connector is the right fit. That seems to be used for time of day, not time duration (Rule Machine Time Variables). Or is that not your experience? I've never actually used the Time Variable Connector.
There is a TimedSession capability that might work - just not sure how that's going to work in this case since the time remaining gets updated pretty often with each poll.
