Queries can all of a sudden slow down, when things were all hunky dory. This is the biggest challenge that developers and architects face.
Getting the data that could establish the root-cause of sudden changes in behaviour in real-time is difficult. Here's how Acceldata proposes the response to such a situation.
Please note that this method of debugging slow queries/jobs applies for the following stacks.
- Hive Tez, MR, LLAP
- Hive on Spark
The section highlighted in Yellow provides sufficient information about runtime parameters and configuration parameters if they have changed