I'd drop the slow query limit to 5 from 10, then enable slow query logging for a while.
Max memory limit is configured to be a gig more than the server has installed in it.
This is troubling:
Current join_buffer_size = 1.00 M
You have had 10940 queries where a join could not use an index properly
You should enable "log-queries-not-using-indexes"
Then look for non indexed joins in the slow query log.
It would be really nice to know what query(s) cause this and what indices should be added.