batch: Improve distributed query engine.

[liurenjie1024]

Improve distributed query engine according to this design: https://singularity-data.quip.com/DMRPAbITntNc/Distributed-Query-Processing-Engine-Design

Query manager(frontend):

• batch: Correctly handle task request failure #2422
• batch: Correctly handle task execution failure. #2423
• batch: Export metrics for monitoring. #3830
• Export execution graph status to dashboard. #3831
• batch: Use determined test to test query manager. #4051
• batch: clean up query manager status when query done #5028

Task manager(compute node):

• batch: Monitor compute task status. #3834
• batch: Add metrics for tasks. #3832
• batch: Monitor memory usage of executors. #4670
• batch: Add metrics for task manager. #3833
• batch: Monitor task pending and execution wall time of tasks. #4700
• batch: Create executor level metrics. #4844

Execution:

• batch: pipelined table scan #1870
• batch: Execute root fragment task in frontend. #3269
• batch: True parallel table scan. #3237
• batch: cancel running task when received ctrl-c. #5027
• batch: clean up task status when task done #5029

Optimizer:

• Tracking: Implement view matching algorithm. #284