As noted in comments, some of this is dependent on genre. As Albrecht notes in another answer, military music made use of drums and sometimes other percussion in all periods. Same thing with ceremonial music that often had militaristic ensembles playing at it. It seems like dance music for popular use also made frequent use of percussion.
However, percussion is very rarely indicated in any early music. Renaissance performance groups today frequently use it when it is genre-appropriate (e.g., for dances, etc.) because we have iconographic evidence of percussion (i.e., paintings and drawings of performances with people playing drums, etc.), not because there are any notated percussion parts.
In general, many people may be unaware of how fragmentary many of these early sources are. In many cases, renaissance dances may only have an extant melody in the actual written source, which performance ensembles use to generate a piece for an entire group of instruments (in a similar way to how a melodic lead sheet might be interpreted by a jazz ensemble today). What we hear on performances of renaissance and medieval music often contains a huge amount of speculation as to what they may have actually sounded like. In other cases, we may have a written melody in one source and then an intabulation for lute or guitar with harmony in another source (as these sorts of arrangements were quite common), and the two are used together to generate a kind of "lead sheet" with melody and harmony for performance today.
David Munrow in particular -- while a leading figure in the modern "early music revival" in his brief life -- was well-known for inventing all sorts of nonsense in his "recreations," which are probably far from any sort of "authentic" performance in many cases. (See here for a lot more info.) And while I love Jordi Savall (and have been to a few concerts), his performances too tend to resemble a sort of "jam session" that sometimes has authentic connections to historical performance practice and sometimes is mostly pure modern invention. The thing about such performances is that they do draw audiences, and often have some sort of grounding in the type of ensembles and techniques used historically. Whereas a literal performance of only the surviving original source (like a single melody) wouldn't be very satisfactory.
Anyhow, back to the central issue of percussion: it existed in dance music throughout history and never left. In other genres, it never existed. For example, it would be completely inappropriate for percussion to accompany a renaissance mass (other than maybe ringing of bells at an appropriate service moment), and that remained the case for most sacred music throughout the baroque period.
Part of the problem is that we have so few sources for instrumental music at all before the baroque, so ensembles do the best with what we have, which was often for things like dance music. That sort of music continued to be written down in later eras, but we tend to think of them as "folk dances" which still only had a notated melody or whatever in the baroque and classical (and even romantic) periods, so they perhaps aren't as frequently performed today, even if they were likely accompanied by percussion historically as well.
Bottom line is that the question's apparent trend is merely due to selection bias in what sources we have for what sorts of genres vs. what tends to be performed today. Well, that and the fact that the music we tend to emphasize in "classical" ensembles from 1600 to 1900 often is associated with elite music played at court functions (i.e., fairly small, intimate venues where percussion was out of place), churches, and then in concert halls, where loud boisterous drumming was viewed as less "refined" and unnecessary compared to the music at the local dance hall.