Integrating the changing current over the draining of the capacitor is correct and quite fun, but there is an easier way: use the energy formula which already has that integral done for you.
Calculate the energy stored in the capacitor, at \$V_a\$ and \$V_b\$.
$$ Energy = 0.5 \times C \times V^2 $$
Subtract the final from the initial to find the available energy.
Divide this by the power drawn to find the time it will take.
Last thought: the boost converter efficiency will drop as the voltage drops. Most of the energy is delivered while the capacitor is fairly full, so you can assume a constant efficiency near \$V_a\$. To get it more accurate, you might need to try that integral after all.
The OL is supposed to tell you that you have overloaded the multimeter in some way. If you look at the datasheet or the specifications of the multimeter you linked, you will notice that it has a capacitance range of only up to 100µF. So the capacitors you are trying to measure are too big (if they are good).
Another overload condition is a short circuit of the leads, it can be imagined as a infinitely large capacitor.
A note on the autoranging with overload conditions: it may take more than 3 seconds to get to the right reading as the multimeter tries the actual measuring range, detects a too big value, switches to the next, detects an overload, until it is at the largest range. I just tried this with a top range Gossen Metrawatt Energy and it took roughly 5 seconds before I got a reading on a 22µF capacitor. Before it would display OL. My HP/Agilent/Keysight 34410A takes around 8 seconds to get the reading, but the display freezes before displaying something.
When you are measuring capacitance with normal leads you might want to hold them as steady as possible and not closely parallel together. Another way is to tape them together and use the Zeroing function of the meter to cancel the lead capacitance out. Additionally you should try to remove your hands while measuring as your body will have a significant effect on the measurement.
A defective capacitor might end up at 30nF, I've taken apart the power board of my LCD monitor and there were some really bad caps, also specified for several hundred microfarad and now in the nanofarad range. Another good indicator for a defective capacitor is the increased ESR (equivalent series resistance), but only specialized LCR meters will give you that value.
Best Answer
I doubt you have bad caps. 100uF is a lot, the meter could just not have the resolution to read the amount of charge it's pumping into the cap. If you have a function generator handy you could measure the capacitance by applying an AC signal to the cap in series with a resistor and see where the voltage amplitude falls to 0.701 of the original value (which is your RC pole).
In terms of what's wrong with your meter, it might just be too large. Try testing on increments of larger caps going up from 30uF.