I think it really depends on what type of advantage players gain by doing this, and how much rolling multiples is perceived as necessary to focus on particular types of content.
It’s really not a bad thing that a particular player might want to play a second version of their favorite class, for example, or that they might gain some benefit from doing so. This is essentially what happens already with alts. There are many advantages to having geared alts, which is why players like having them.
Also, players at the level of Method and Limit already push this sort of thing to the extreme. If having multiples doesn’t require a massive amount of grinding, then cutting edge players wanting to have a second Covenant for their primary class isn’t a huge issue, imho.
We saw this in Legion with class-specific content. There was a pretty big incentive, even for casual players, to level and play one character of each class, or as many as possible.
In Shadowlands, there might be more incentive to double up on your favorite classes. I’m thinking of doing it with warlock, paladin and warrior, which are the classes I play the most. But that’s not because I’m concerned about being “optimal,” it’s more just that it could be fun to take them down different progression paths.
Obviously, there will be a lot of players who collect all the appearances, or as many as they can. So I think it comes down to how the total package offered by each Covenant is designed. If it ends up feeling like one Covenant creates a character that is good in raids, for example, but terrible in dungeons and pvp, that would most likely represent a design failure because the game as a whole encourages crossing over.
It shouldn’t feel like I need one paladin for raids, one for mythic+ and one for pvp.