So we are having some issues with production typesense. 255GB dataset. It was running ok for a few hours, good amount of traffic, but then typesense starts pegging all CPU cores at 100%. So brought down search, stopped all searches from coming in, so there is zero requests going into typesense, but even after 5 minutes it's still pegged at 100% max CPU. All searches are being performed with a search_cutoff_ms of 2 minutes, but after 5 minutes of no seaches, the DB is still hung at max CPU usage on all threads assigned to it (96 threads on a 120 core system in this case)