Update target MRs for changes and commits to be more realistic

While reviewing other tests it become obvious that the current large MR we're testing against isn't as realistic as we'd like in our test data for testing Commits and Changes endpoints. On further review it looks to have been an MR open by mistake trying to merge master into an older branch, therefore creating an unrealistic MR.

New MRs have been found in our existing test data that is more realistic so this MR changes to those.

The new MRs are different so the numbers change slightly as well but the current open issues won't need to be changed as we're still in the same ballpark. To be clear the performance issues we've found still stand, we're just moving to a more realistic MR to keep us right moving forward.