Single stage charging at bulk charge rate is rated at 2.55v per cell x 6 cells = 15.3V. That is if you can make a 10 to 12% amp charge rate of the 20 hour rating with your solar panel... and it's rated at 105 ampere hours. So you need to make minimum 10.5 amps to 13 amps, constant current, at 15.3 volts. Your amps should taper off to 3%, or about 3 amps, at 15.3 volts.
That's with no load on the battery... add in all your constant draw with refrigerator and other goodies, and it's eating right off the top of your voltage and amps going in to the battery. So you either need to make more amps, or push harder with more voltage, to offset the constant parasitic losses.
Basically, you need a panel set up that will make 18 to 20 amps, at 15.3 volts, to give you enough amps left to cover the parasitic losses from the refrig and other goodies going off and on during the day. That is so that you have the 12 amps left to put into the battery, and the moment you start using items while recharging the battery, you get voltage drop also, to the battery, interfering with the charging, unless you get more amps.
Perhaps now it is not a surprise why your 31 is acting up... You don't really have a 3 stage charge controller, not with clouds and other stuff blowing by. How many amps a day do you get going into the battery? Something tells me you are coming up way short with all the amps you're using per day, in the voltage department, and in the amp hours department, compared to parasitic losses. Something is not being accounted for properly.
Then, from your charge controller, to the battery, what are your voltage losses, and voltage at the terminals vs what is programmed into your charge controller?
Something is really amiss.