If the chunk migration is in the "steady" state(data cloned/oplogs fetch done and some potential writes happening on the source) and has a slow session history migration before committing, we could flood the Data Transfer Progress logs.

Here If the chunk migration is in a steady state, the conditional check for "one second" wouldn't occur and the source shard will be sending _recvChunkStatus commands to the destination shard in an unthrottled manner. https://jira.mongodb.org/browse/HELP-65911 led to 26k messages in 23 seconds.