I’ve ran recently to a problem. We have a MySQL 5.7 Innodb replica with RBR replication (8 slave parallel workers). Now, until this Monday, our master executed 2500 inserts per second (in average), and the replica seemed to be fine with that. But recently the number of inserts went to 2800 inserts per second, and the replica does not seem to catch on (at evenings maybe). I checked the slave status. It seems like the IO thread seems to work just fine (there is no problem with getting events from the master), but the SQL threads don’t run that fast. So the question is - how can I improve that - increase of parallel workers does not seem to help at all.
By the way, here is the result of sql show status: