We have a DSE Prod cluster with DSE 6.0.10 version with "9 nodes Spark-Analytics mode as DC1" and "5 nodes Cassandra mode DC2 ". We have multiple SPARK-SCALA Jobs scheduled to run every 15 min as 1 time CLIENT mode and 4 applications with SPARK-STREAMING running as CLUSTER mode with PROD cluster, to give you some back ground. Please see below findings from OpsCenter to troubleshoot DSE Keyspaces and checking how to maintain DSE keyspaces. Your insights helps us, thanks.
TOMBSTONE Count:- dsefs.transaction_log (20479), dsefs.block_offsets (20479), dsefs.names (5119)
WIDE Partitions :- dse_analytics.rm_executors (2741 mb)