I will personally be voting do not ban on Dragapult. Many times have I considered flipping over but after a fair amount of playing in this metagame I have noticed a lack of usage of other mons that can reliably answer Dragapult without getting too heavily punished.
Player skill differential:
One thing that I particularly like about Dragapult in this metagame is that it tends to promote careful analysis of team structures and play to be used to its best potential. This to me is exclusively a benefit - for the most part I believe that a good metagame should be able to reduce matchup to a bearable level whilst maximizing skill in play. Dragapult ends up fitting these requirements for me as the checks individually to all of the sets are quite fast but because of that careful team analysis you have to know when you could be stumbling into the wrong variant. But for most checks, this is not too major of a punish if they get the prediction of DD Z move wrong. For some checks, the opposite can be devastating (for example, Blissey) but to me this is an inherent risk vs reward with using Blissey as the main Dragapult check - and for the most part, it isn't too threatening if you get this prediction wrong given the right team structure as it will likely only take 1 kill at best.
The "kill" button:
One major argument I see about Dragapult is how getting DD Z's play wrong will often result in a sweep. I actually think this tends to go with a lot of good Z move users - Magearna, Kartana - even some lesser variants like Hydreigon and Cinderace can easily leave you in the wrong position if you predict the Z move wrong. One may mention Dragapult's insane speed as a way to refute the revenge killing argument, however, one thing that isn't mentioned is how a lot of Steel-types can revenge kill Dragapult anyway due to the lack of coverage it has for them. Heatran, Ferrothorn, Magearna, Magnezone and the like can all check Dragapult after its Z move has been burned just the same way revenge killers like Cinderace would be able to check Magearna after its Z move has been dumped. Whilst this isn't air-tight as an argument, and leaves a lot of loopholes to promote that potentially all of these Pokemon could also be broken, I do believe this is something to consider quite heavily.
Pult's presence in the metagame, and how its partners can seemingly push it over:
Be warned - this part of the post is slightly rambly. If you're looking for arguments that are specific to not banning Pult this is probably not the part of the analysis you are looking for, this goes more into depth about how the other partners are broken alone and how Dragapult abuses them (aside from Rillaboom).
Cinderace:
These two in particular actually form a really frustrating core to deal with. Because of how broken and variable Cinderace is, it can pair multiple sets with multiple combinations of Dragapult and force out other good answers to it. 4 Attacks Cinderace can form a really annoying dual pivot core with Hex Pult - and there are multiple different variations to boot, like Triple Death Move (Pyro/HJK/Gunk) Cinderace + Electrium Pult, and GunkZen Cinderace + Twave Hex Pult. However, one may instead question what individually one of the two tends to do, and for me the most overbearing parts land on Cinderace for being an insanely restrictive mon due to its amazing coverage moves, and pairings with almost any broken on this list, and even some outside of the list that Pult would not pair well with either, like Slowbro, Tapu Koko, Hydreigon, etc.
Mega Latias:
These two actually synergize very well on paper but I generally tend to feel like due to the conflicting natures which DD Z pult and Mega Latias exhibit, teams with both of them generally tend to feel like they are gasping for air when revenge killing is taken into account. Other pairings like Hex Dragapult do tend to work much better but tend to rely on the opponent bringing specific answers like Mega Tyranitar and overloading them without the opponent being able to win first. For me, I yet again feel like a lot of the pressure done by this is caused by the restrictive checks that Mega Latias has and its ability to switch up coverage moves and completely change its checklist is what makes it the more broken partner.
Rillaboom:
Rillaboom's synergy with Dragapult is a very awkward one. Indeed, you could argue that this isn't on the same level as the other partners, and I'd intend to agree with you on that. What this does exhibit is how powerful these two are as support for HO, and occasionally general offensive styles, too. DD Z Dragapult + SD Rillaboom is a very powerful and common core on HO due to the fact that both can overwhelm some Rillaboom checks very easily, and Rillaboom just so happens to be capable of outright eliminating some good Dragapult checks for it. The only issue is when the Dragapult check is something like Knock Toxapex or Clefable, then the two just don't connect very well. And that often happens to be the case with a lot of these Dragapult cores - they tend to fail under a lot of matchups.
I'd like to share some opinions of arguments I have from this thread.
Punishing Passivity in the Metagame
The final point I want to respond to is this notion that Dragapult thrives in the current metagame in a way that is unhealthy. I would argue strongly in the alternative that punishing passivity in a tier where Clefable, Toxapex, Blissey and Mega Latias are S-tier in viability (and even Slowbro(!!?) is getting mentioned for a suspect test) is actually a very desirable and healthy quality. I am absolutely not a stall-hater, and obviously Dragapult isn’t the only option we have to punish these mons. However I don’t see why punishing S-tier passive mons is considered an unhealthy trait, while these mons being solidly S-tier to begin with is not even remotely a cause for concern going forward, or the fact that apparently over half of the people who achieved reqs to vote Dragapult out did so using at least two different variants of stall itself. Clearly passivity is thriving in the current metagame and the fact that Dragapult punishes that is to me healthy, or at least doesn’t seem to be a debilitating problem.
For the record I actually think that Dragapult is pretty poor at punishing passivity. What it does best is punish offense - most balances will often have hard checks like Knock Off Toxapex or Clefable that can easily scout most variants and can at worst deal significant chip damage to a perfectly played DD Z set. However, the checks on offense are slightly limited. For the most part, the checks I have been using on offense are Melmetal and Mega Tyranitar, however, Magearna can also work given a heavy amount of Defense investment, and other checks such as Mega Scizor can pull off the assumption quite well. The punishment from other sets like Specs and Hex can be pretty heavy to these partners, however they are not heavy enough to be game-ending either. That being said, the playstyles of each also tend to be different, and for me it is easy to tell whether a player is playing DD or not, as DD is pretty poor as an overall Pokemon at forcing things out if the Z move is not burned. That being said, I do agree with the other two parts of the argument, I just believe that the idea of punishing passivity in the metagame tends to be a strawman argument that can take away from actually relevant points.
Also, I just thought of this a day later, but I can't help but notice that a lot of the anti ban sentiment revolves around common cores that check pult. Never a single mon always two or more.
Because that's what we need more of: offensive threats so potent it further prevents stall users from actually managing to fit an offensive wincon on their team and increasing net suffering for all involved. It doesn't discourage or punish stall, it just forces them to run even more fat in preparation.
I don't really think this is the case with Dragapult - checks are plenty enough to every set such so that stall will not have problems accomodating for it. Stall cores have barely changed in adaptation to Dragapult - maybe the occasional Knock Off Toxapex exists but the amount of Unaware users and easy pivots that exist on there already don't really effect stall's supposed inability to fit a wincon. That being said no wincon stall sux, but will always exist unless you have so many strong offensive threats that it is practically impossible to wall everything without revenge killing (see: SM metagame). So I don't really think Dragapult is doing much in the way of making stall more defensive than it already was.
These are the thoughts that I have had to come to the conclusion of voting do not ban. But this is just my overall opinion, and so some points don't make too much sense in the context of a suspect test. I just wanted to get my thoughts out.