You could have a compressor on your main mix, but mostly for artistic effect rather than speaker protection.
I prefer to compress channels, which is why I spent what I had to to get 16 channels of compression of my 16 channel board. It always seemed that if I had 2 channels of compression I needed 3 or 4 , and when I had 4 channels available I need 7.
With only two channels you'll have to prioritize. Just keep the insert cables from the compressor handy and pop one into the insert of whatever channel seems the most dynamically out of hand. Have them set to the highest threshold, 2:1 ratio, medium attack and release and no makeup gain. After inserting bring the threshold down until there's gain reduction on the meter. Now you have to use your ears and decide what parameter adjustments are needed.
Threshold is the dividing line between the louder parts that get compressed and the quieter parts which don't. Set it high enough to leave some quieter parts of the signal unaffected.
Ratio is how much the loud (above the threshold) parts are affected. Higher ratio means more less change in level.
Attack and release let compression happen slower or faster. Attack affects the response as the signal rises, letting more or less of the initial transients through. Release affect the response as the signal falls which can prevent unnatural sounding effects.
Makeup gain lets you add gain after making the louder parts less loud. For live sound I'd avoid this on anything with potential for feedback unless I could set it at sound check. Just leave it set to unity.
Be careful about applying compression to vocals if you're mixing their monitors. Singers can stress their voices when compressed in monitors, and it makes feedback more likely.
