Select_scan and next-row-request rates are way too high, so I agree with Øystein, many queries are requiring table scans, you need to identify and optimise them.
It seems to be rotating the innodb log file every minute, that's probably slowing you down, optimum is usually about 30 mins, so consider increasing innodb_log_file_size, and see
https://www.percona.com/blog/2012/10/08/measuring-the-amount-of-writes-in-innodb-redo-logs/
3k inserts/sec is busy, see
https://dev.mysql.com/doc/refman/5.7/en/insert-optimization.html re optimising inserts.
Use free -h, or top, to see if the system is swapping. If it still is after fixing the inefficient queries, you'll need to identify the queries that induce swapping and fix them or adjust other params to get rid of swapping.