The +2.1 dB was found during the intro cutscene. A peak from the music intersected with one of the 'ping' cursor opening sound effects in just the wrong way. I have software for measuring desktop audio in LUFS for my audio-related game dev work.

I want to preface that my knowledge in this area is solely from my experiences for the purpose of game development. I am not a professional music or audio engineer. This is an amateur opinion.

As far as I can tell, all of the music is as loud as it can be without clipping. When no sounds were occurring, the level tracks hovered between 15, 8, and 6 LUFS, respectively, and peaked near -0.1 dB. For the purpose of listening to this as music, this loudness is desirable, I get every little detail. It's just that's it very loud for a video game.

The moment any sound effect is added onto one of the peaks, such as the intro sounds or the player shooting, it goes near or above -0.0 dB. There's no room for any additional sound to be added on top of the music. This is why music in-game tends to be quieter than on an original soundtrack.

It goes beyond -0.0 dB regularly in the intro, as well as levels 1 and 3. It doesn't occur in level 2, because there's no sound effects. It's not the most egregious clipping; I've heard much worse and much louder, especially in game jams.

Considering how quiet most of the sound effects already are, it probably wouldn't take much of a reduction in music volume to amend this. The important thing is to make sure to measure the final volume of a game's music and sound effects playing at the same time.