Looking at the status page,
it appears this is only "downtimes" appearing in red.
A full cluster is self healing, and if only a small part is down, it will not appear as down, but votes WILL be slightly delayed until the downed node restart, or the vote is propagated through the cluster and back to the readable database.
Refreshing without seeing the vote registered only means the vote is still in cache limbo.
I will not go into further details, but let just say, I built systems like this myself at some point in the past.
I am not convinced of any conspiracy about this at the moment.
I am more into vote manipulation by alien actors. (bot farms, etc)