How did you measure the coil current? What coil current did you measure and what was your Vref voltage? Was your stepper motor stepping while you measured the coil current? What is your power supply?
The reason that we suggest you measure the coil current while in full-step mode is because that is the only mode where the coil current magnitude is the same for both coils and independent of the step you are on. When microstepping, the coil current is given by:
coil 1 current = current limit * sin(step*2pi/(#microsteps*4))
coil 2 current = current limit * cos(step*2pi/(#microsteps*4))
The magnitude of the current vector is given by
sqrt[(coil 1 current)² + (coil 2 current)²] = current limit
So what you end up with is a vector of constant magnitude that is just rotating with every step. This keeps the motor torque constant. But the current in each coil is changing with every step. Unless you know exactly what step you are on and feel like doing the math, it is difficult to work backwards from what you measure to what you actually have the current limit set to. The DRV8825 datasheet has the step table at the end, so you can see what the driver is trying to limit the coil current to for each microstep. When microstepping, it is important that your current limit is set to the rated current limit of the driver, since on some steps, 100% of the current limit will be flowing through one of the coils (the other coil will have little to no current flowing on these steps). When in full step-mode, you can actually set the current limit a little higher than the rated current, since there will always be approximately 70% of the current limit flowing through each coil. Does this make sense?
The motor temperature should increase with increased current, since the heat is the result of power dissipated by the resistance of the motor coils. P = I²R, and R is a constant, so if P is going down, I is going down. However, the current is a function of more than just the current limit setting. For example, the faster your motor is stepping, the lower the current will be through the coils (assuming all other conditions are held constant).
I don't have much to say about the ramp values or ramp time. I think you should just try playing around with the different parameters to see how they affect your results. It sounds like you are changing the step frequency discontinuously; you might get better results if you ramp the speed up more smoothly.