The causally_consistent_hedged_reads_jscore_passthrough, causally_consistent_read_concern_snapshot_passthrough, and sharded_causally_consistent_read_concern_snapshot_passthrough tasks have all been observed to fail due to an excessive number of resmoke.py jobs (--jobs=10). The failures have been a mixture of slow resmoke.py logging and having too many files open on the system. Reducing the number of resmoke.py jobs should address both of these issues.