There are two issues here: 1) whether sleep clause is itself outdated/incompatible with modern tiering principles (a claim which may affect sleeps present in future gens); and 2) whether sleep is broken in current Gen IX OU, even with current sleep clause, or if it might just be darkrai that is the problem.
1) Should sleep clause be removed entirely on principle, including for future iterations of OU?
^This train of thought is the nail in the coffin for sleep from a self-consistency standpoint, in my opinion. If sleep were a new status, there is absolutely no way we would go out of our way to form a complex ban to keep it around. Sleep clause's blatant contradiction with modern OU tiering philosophy has bugged me for a long time. The fact that it's current implementation requires deviation from cart mechanics is even worse. If it were up to me, I would just get rid of (non self-inflicted) sleep for all current gen OUs going forward.
That said, as much as tiering is based on clear principles and the logic that flows from them, it must be acknowledged to some extent that tradition/nostalgia plays a big part in deciding what things are allowed. To that end, while I do not like sleep clause at all, I do have sympathy for people who would like to keep it around because the availability of sleep moves in some form has historically been a part of competitive singles. If a large part of the Smogon community as a whole thinks that sleep clause is worth keeping solely out of tradition, I don't think it would be wise to remove it for all time, even if I dislike it. For the same reason, I want to be clear that should this line of thinking become policy, I think both I and most others would NOT want this to be a retroactive change to prior generations (this is obviously just an Gen IX thread but when making broad changes like this often eyes soon turn towards making old gens consistent with modern ones). There is no need to go overturning other settled generations to fit with our modern conceptions of what tiering should look like.
To touch on some sleep related odds and ends:
Dire claw can go with sleep, no need to make an exception (I feel likewise with relic song for consistency tbh but no one cares about meloetta so whatever). I do acknowledge that yawn has a more competitive flair than the rest of sleep moves, so I understanding wanting to separate it from the immediate sleep moves (you could imagine a ban on moves that cause instantaneous sleep), but at the same time I also think that the preference for less complex clauses in battle is a good one. So I can go either way here. Now what to do with effect spore causing sleep is a whole other issue; thankfully, right now there is no pokemon which has effect spore as its only ability, so I think it is fine to ban the whole thing, but one could imagine a future where an otherwise balanced pokemon with only effect spore getting booted from OU simply due to its ability (getting frosslass vibes here); I will leave that to future players to deal with this should the problem arise.
2) Setting that aside, is sleep as it currently exists broken in Gen IX OU? Or is it Darkrai?
This is a tough call. I will start by saying that in general, I do NOT think that sleep cause in its current implementation is inherently uncompetitive, in general (although how it applies to hypnosis in this generation in particular will be discussed later) There is obviously a large element of luck in sleep, but even so several previous metagames have managed to remain skill based and competitive, even when sleep is a big part of the ecosystem.
I think generally sleep's power is greater in more offensive metagames, where it is harder to recover from the instant loss of a pokemon due to sleep, and the momentum gained by putting a pokemon to sleep can be immediately capitalized upon. I think the thing that sets Darkrai (as well as to a lesser extent Iron Valiant and A-Ninetales) apart is that these pokemon get you so much immediate advantage from a 60% roll that it gets to a point where the outcome of one or two attacks can be disporportionally game warping, leading them to be uncompetitive. As a result, I do think it makes sense to do something about this mechanic.
On the flipside though, in high power environments, the downside of just wasting a turn due to the low accuracy of most widely distributed sleep moves similarly becomes harder to justify. To that end, I am not sure if hypnosis darkrai (if it is allowed to exist as it does now) will ever take off consistently in a tournament environment; even when the roll is in your favor, missing a hypnosis can be a big risk when you have a lot riding on the line. This is not to say that I don't think that these sets will be used at all in tourneys (there are already some successful tourney results from hypnosis darkrai), but I suspect that should they be allowed to persist, in the long run these sets would fade out of popularity due to inconsistency (in contrast, in a ladder environment, where high variance is acceptable if the expected value is large enough, I think darkrai would remain a scourge). As a result, in the tournament metagame, I think ultimately it is spore amoonguss that will prove the most annoying. However, in the current moment, I doubt amoonguss alone is what is making people call for action on sleep. Iron Valiant and A-Ninetales are similar; I do think they have problematic elements with regards to sleep, but without darkrai I am not sure this issue would have generated as much attention.
On that note, I do think that in this particular instance, it is very difficult to separate darkrai from a total sleep ban. Is sleep clause really just largely uncompetitive in this gen, or is darkrai the one who is really pushing it over the edge? If I am trying to be dispassionate, I would maybe lean a little towards banning darkrai first before going all out on sleep (provided there is not a fundamental alteration towards the tiering policy on sleep, which I would prefer). I do think there is a good chance that sleep may remain uncompetitive even so, but there is no harm in revisiting this down the line and possibly freeing darkrai in exchange later. It just feels unfair to hit all sleep across the board when in this moment it is very unclear where the real source of the problem is. On the other hand though, I just don't like gimmicky sleep stuff lol (in the context of this gen), and if I were redesigning the tier to my own wishes I probably wouldn't include it, not to mention the fact that banning darkrai may just be delaying the inevitable. So ultimately I'm ok with either path.
TLDR
1) Sleep clause as it exists should be abolished in my view; it is bad enough that it conflicts with modern tiering philosophy, it is even worse that it breaks cartridge mechanics. However, I understand the community at large may feel some traditional attachment to sleep as it has been implemented historically, and I think it is ok to violate consistency for the sake of maintaining the status quo, provided that is what the majority feels (once again, if it were up to me, I would kick sleep out altogether
2) If we are not going to fundamentally rework sleep clause, and are just concerning ourselves with whether sleep clause is broken in Gen IX, I do think there is an issue with sleep that darkrai has brought to the forefront, but it is very difficult for me to say definitively whether it is sleep broadly or darkrai specifically that is the problem. I think the most "objective" option is to ban darkrai first, and then revisit this later if sleep remains an issue, but my personal feeling is that I do not enjoy sleep broadly right now and would not mind seeing it go entirely for this generation, so long as most others agree.
