The "Analyze Job" feature has MapReduce execution flow analyzer, is a unique solution that gives the cluster level correlation of data with execution flow inside the code. It can bring down hours of execution logic debugging trials by identification of root causes within minutes. User may apply regex validations or user defined validation classes. As per the applied validation, Flow Debugger checks the flow of input data tuples essentially <key, value> pair data for each mapper and reducer in the submitted job.
- Jumbune provides a comprehensive table/chart view depicting the flow of input records through the job.
- The flow is displayed at job level, MapReduce level, and instance level.
- Unmatched keys/values represent the number of unexpected flow of key/value data through the job.
- Debugger drills down into the code to examine the flow of data through various counters like loops and if-conditions, else-if, etc.
- MapReduce job profiler module gives the correlation between the individual execution phases, resource consumption and throughput, for identification and rectification of bottlenecks.