Github user roshannaik commented on the issue:
https://github.com/apache/storm/pull/2203
@HeartSaVioR to see the perf impact on throughput, you want to see what is
its peak ability by setting a high rate... then see what is the highest rate
each binary achieves. Here both seem to be able to easily achieve the 50k rate
and thereafter the spout begins to throttle.
The latency impact is not recorded in these readings... it is key IMO.
At a glance: The memory impact seems a high. The CPU has utilization and
gc activity has also increased... not sure by what %.
---
