Your RTC has a BBRAM in it. I've done extensive testing on DS1230Y BBRAM in combination with various types, but particularly Maxim/Dallas DS89C4x0's.

I ran a 10 million-cycle test using RC reset and detected not even one BBRAM corruption, whereas, with a relatively slow Vcc rise time, using a DS1232 supervisor I found that I had BBRAM corruption more than 10% of the time, though that observation was not rigorous. I'm of the opinion that the problems occur during the decay of Vcc, as I have observed the MCU apparently accessing the BBRAM during active RESET (from the supervisor) during Vcc decay. This apparently occurs during the window between when the supervisor asserts reset and Vcc falls below the valid level, in this case 4.75 volts, but is not yet low enough to cause the BBRAM to inhibit writes.

Pay no attention to Erik Malund's comment, as he doesn't even know what the rise and fall times of his power rails are, nor does he know what their potential effects are. He lacks the interest and diligence to figure it out. He just likes to complain when his assertions are challenged.

There's a problem with your circuit, though, and nobody can help you without knowing exactly what it is. Presumably, you've thoroughly studied the datasheet of your MCU and your BBRAM/RTC, so you're expert in the details of their operation. If that's not true, it would be best to study those documents carefully.

RE