A steady native thread leak has been observed under a sustained query mixture on the followers combined with a query + update mixture on the leader. I have not determined whether the problem could be replicated with just BSBM Query on the followers or if it depends on the concurrent updates.
The root cause appears to be distributed GC. The following issues [1,2] also have this as their root cause:
 https://sourceforge.net/apps/trac/bigdata/ticket/433 (Cluster leaks threads under read-only index operations)
 https://sourceforge.net/apps/trac/bigdata/ticket/437 (Thread-local cache combined with unbounded thread pools causes effective memory leak)
 https://sourceforge.net/apps/trac/bigdata/ticket/677 (HA deadlock under UPDATE + QUERY)