1. rewrite metric module, add disk capacity metric 2. add Cluster load statistic proc to observer cluster load balance status 3. fix bug: show table status from db throw Nullpointer exception 4. performance: change push_down table size to 1024 in HashJoinNode