Unfortunately, I see my weaviate container instance stuck at CPU intensive cycle for doing tombstone cleanup again, and this time it isnt due to shutdown corruption. I noted that this round, I have caused it delete over 36k in ~3min span.
I will take a look at TOMBSTONE_DELETION_* to see what I can tune. But it does appear this is prune to happen if I try to delete a large # of docs within a short span of time.