IIRC, you tend to record unamplified stuff, making it a bit more difficult to set levels appropriately given the broad dynamic range between quiet passages of music and thunderous applause. Basically, you have three options:
<1> ride the levels (ugh)
<2> set levels for the applause (probably very quiet music)
<3> set levels for the music (applause may clip)
I hate doing <1>. It's like a rollercoaster ride, detracts from my enjoyment of the performance, and results in rather challenging post-production to smooth things out. I've bounced back and forth between <2-3>, but have mostly settled on <3>, damn the applause. It's the music I really care about, after all. Then, I'll either compress / limit the applause in post, or if the applause doesn't take to compression / limiting very well, just remove it outright.
As for what target dBFS levels, well...that depends on your gear and whether you're running 16- or 24-bit. The non-specific answer: set 16-bit levels so your music (assuming you're using method <3> above) peak approaches, but does not hit, 0 dBFS. For 24-bit, you have some more headroom, so you need not approach 0 dBFS as closely as possible (without clipping). I typically aim to peak -12 dBFS when running 24-bit. But ultimately, it depends on the source material, recording gear, environment, and how loud I expect the occasional music peak to be.