I think Phil's point is very important. While I find github/gitlab to be a nice interface for small projects, I'm not at all convinced that it adds much for larger or more complex projects. As an example, the bug/issue tracker really doesn't scale well once you have hundreds of bugs/issues.
Once you have a project with more than ~10k lines of code, multiple branches and large numbers of contributors and bug reporters, administration becomes a big issue. I suspect that few except perhaps casual users will really use the UI's provided by github/gitlab, preferring instead to use just git. At that point, where the repo is hosted becomes less relevant.
My experience with github and gitlab issue tracking has been that you need to either use a different tool or develop custom workflows using the API in order to manage large numbers of bugs effectively. Again, the web interface is OK for casual users and smaller projects, but lacks the level of sophistication necessary for larger teams and larger bug numbers. I'm not saying that the current bug tracker is great, but I've found few bug tracking systems to be great and the current one is adequate once you invest time into it.
I'm certainly not against something like gitlab or the gitlab interface, but I don't think it will be as beneficial or 'revolutionary' to Emacs' maintenance or level of contributions as some seem to feel. In most ways, it is just a little bit of UI sugar and a community is more than just a bit of sugar.