Well I've been busy with other things. After talking to other people I'm still thinking I'll use the ESP8266 in the sensors, but I am probably going to build a head to head ESP8266 vs. Arudino Micro plus NRF24L01 and see how they go against each other.
I also manage to process some log data from across last weekend. As usual it seems obvious with hindsight, buy very obviously having the temperature sender inside the plastic box with the electronics acts as a bit of a glasshouse, and you can see the temperature spiking up to 35C on a day when outside never got past 25C. This seems to be actually interfering with voltage regulation interestingly, the output drifts as high as 3.6V (unloaded, still) once the supercap is fully charged. On the plus side, even once I hook up the ESP8266 it looks like the 22F capactor _should_ hold enough charge to do what I need overnight and still recharge on an overcast day, although the jury is still out. I'm intending to hook that lot up on the weekend, as I finished a first cut of the sensor software, which itself logs its own VCC and a DS18B20 using an ESP-01.
I cant remember if I mentioned it, but I am also playing with designs for a solar MPPT charger front end although that may be gold plating if things work OK as is.