The esp32 high level source current rating is 40 mA (typ). The Songle relay SRD-03VDC-SL-C says the typical coil current is 120 mA. The board you're using with the relay mounted on has one 3 terminal component that is possibly a transistor or mosfet to help you with power. Also an optoisolator. Try ...

Yes, put the bulk cap on the input to the motor driver. Electrolytic caps have polarity and also tons of capacitance in small space. Think of the motor driver as just redirecting the input DC voltage to the positive or negative motor terminal to control direction (and also pwm to control speed). So ...

You need a bulk cap for the motor V+ supply, try that first. If you look at the L298N datasheet pg 8, it shows a 470uF cap. And in this article, the module he's using has a 220uF cap for each motor. https://lastminuteengineers.com/l298n-dc-stepper-driver-arduino-tutorial/ A larger motor controller w...

The motor and esp each have their own battery, but GND is connected. The motor is inductive, from a changing load or speed it can dip or spike it's power supply. Perhaps it is dumping energy into GND which raises it's potential, decreases the gap between GND and vdd giving you temporary low voltage....

There's an internal pull-up on IO0 and pulldown on IO2, no need for external resistors.
https://www.esp32.com/viewtopic.php?t=5970

Can you share more of your code? Are you using an interrupt function triggered by the counter pulse?

What does ".counter_en = false " do?.

It's 1kHz, 40 divider, why do you expect 2000 counts? Is it run for 80 seconds?

The led is connected to TP4054 CHRG pin. With no battery connected, maybe it is doing the "charge complete / AC present" indicator of 20uA pull down. Are you seeing a bright flash or very dim? https://i.imgur.com/mwLLfeQ.png https://i.imgur.com/yWuAvCs.png Maybe your usb port is doing something like...

Read this post, it was published by espressif, you need to manually measure vref and then input it in your code as a calibration: https://www.esp32.com/viewtopic.php?f=19&t=2881&start=20#p16166 The analog_read function is not applying calibration or anything. So if you use analog_read, you must conv...

Yes as chegewara said the watchdog chip is simpler and less likely to fail and the esp32. So you have shifted the problem from the ESP 32 over to the watchdog chip, you shifted the problem from a more likely failure to a less likely failure and that is a success. You also can have a watchdog chip th...

Two point calibration will correct the offset and slope, thus you'll be more correct at both the low end and high end. Efuse vref can be single point or two point, obviously two point is a little better. Two Point values represent each of the ADCs’ readings at 150 mV and 850 mV. To obtain more accur...