Every time we'd look into this, we'd look into the challenges we'd have to face and see if they changed since last time. One of the problems GitLab faces is that it JOINs all over the place. In addition, while most data can be shared on a group/namespace basis, GitLab also does a bunch of cross-group queries in frequently accessed pages.

To put it differently, try to think of an application that does all the things you _don't_ want to be doing if you're going to (or want to) shard. Then imagine that that's basically GitLab, coupled with the load patterns simply not justifying it.