FYI, I was running “compact” against 3TB collection. Would that cause the problem? Our mongo 4.4.28 replication had stalled few times even we weren’t running the “compact”
This happens on a secondary, and once the replication stalls, the CUP & disk usage go almost nil.