Looking at these new graphs, I think there may be a bug in how Pixelfed calculates new posts.

Also that one misskey instance I never heard of that has 6.7M new posts in May looks sus