The esp in sleep mode consumes almost no power at all, but all the rest can additionally draw a few mA. In the link you have a discussion about what and how much electricity it draws additionally.

http://community.m5stack.com/topic/1162/getting-longer-battery-run-time

It will be best if the sensor is connected 5-10cm away from the device. The sensor connected directly to the connector will be heated by the device through paths and goldpins and you can not help it.

As far as dht11 / 12 is concerned, I already subtract these 2-3 degrees in the program from the read temperature, which are overstated. But you have to check it yourself experimentally if the result is overstated with any other sensor in the temperature range you want to measure.

Unfortunately, I don't have dht22, so I won't tell you anything specific about its accuracy or results.

you should additionaly use start timer block

example:
https://github.com/stonatm/m5stickc/blob/master/mqtt.png

set timer can be use to change parameters of timer which you run by start timer

@fb24 Hi. A few days ago I made a custom block for ds18b20 support. Maybe you're interested

https://github.com/stonatm/UiFlow-custom-blocks/tree/master/ds18b20

You will need the UsbAsp programmer to program atmega. The maximum supply voltage can be found in the atmega datasheet because it depends on these few letters or numbers after the inscription 328

see this file

https://github.com/m5stack/M5StickC/blob/master/src/AXP192.h

It is a part of M5StickC library for arduino

You find it in a functions that help you to determine when power is connected to usb port, what voltage input and what current flow.

You should interest this functions of axp192
GetVinVoltage();
GetVinCurrent();

here you have the library to support ds18b20. You would have to copy these files to your device and from UiFlow call the appropriate micropython commands using the blocks Advanced -> Execute -> Execute code:

https://github.com/micropython/micropython/tree/master/drivers/onewire

Could you try uPyCraft. I use them on windows system and i can transfer files to/from Atom Lite/Matrix with burned Uiflow firmware. Under linux where i use Linux Mint 19 uPyCraft i cannot run it (problem with installing proper libraries needed to run)