question

@nickdb There's a few reasons: it's a small setup with small(ish) batteries (20Ah capacity). Typical loads are small, virtually always less than 0.5 amps, so this capacity is enough for a few days even with no sunlight. Connecting directly to the batteries would eliminate the usefulness of the "BatteryLife" option turning off loads when the batteries get low and not include the load of the Pi from the load output current/power measurements in the dashboard; a smart shunt would be nice, but costs more than the whole setup. I'm using LiFePo4 batteries and, although they have a built-in BMS, I prefer to avoid relying it for low-voltage cutoff.

Also, it simplifies the wiring for me: the batteries are connected to the MPPT controller via fused wires. The solar panels are connected to the MPPT controller. A small external fusebox is connected to the load terminals of the MPPT controller, and the loads fed off of that. Since there's no inverters and loads are very small, feeding the Pi or other loads directly off of the batteries would add complexity and require a bunch of re-wiring.

> In my experience, when someone is prompted "are you sure" they generally are sure, until they suddenly aren't :)

I feel that in my bones.

Even so, having a prompt at least makes it slightly harder to shoot myself in the foot so easily.

In my case, the footgunning was minor and easily resolved, but I can see how it could be much more of a hassle for people with more mission-critical needs or systems less accessible than their shed.