It could have to do with drum mikes.
Remember that aiming a mic at a drum's top head is the opposite polarity from putting a mic inside a kick drum aimed at the beater.
Outside a drum, the impact of stick upon head causes a rarefaction of the surrounding air.
Inside a drum, the impact of the beater moving the head towards the mic causes a compression of surrounding air.
Neither is a right or wrong way to mic a drum, but the careful mix engineer might flip polarity of some of the mics, and thus cause addition of signal, which could potentially explain this.
On the other hand, if the peaks are bass guitar, maybe the battery in the phantom DI box is getting low?