Thanks for reporting the issue!
We do monitor a few metrics on Wikida Query Service [1], including the replication lag [2]. The replication lag seems to be around 10 seconds at the moment (which is more than reasonable). So I suspect we have an issue with some edits that have been replicated correctly rather than a systematic issue. We have had some issues with reloading all data from scratch recently, but that should have been fixed and that was before May 20th.
I'm not entirely sure what is happening here, but I created a phab task [3] to track this. I hope Stas can have a look into it (or at least help me have a look into this).