NCMP Profiling (pending latest perf fix CPS-1171)

Execute  use-case performance profiling (focussed but no limited to registration and module sync use-cases).

we need to understand were the time is spent to allow effective future performance improvements.

Number of  Modules (mocked DMI) 180

detailed (sub) use cases to investigate/compare

1. Register 1 (first) batch of 100 cm-handles with no parallelism (level =1)
   1. include breakdown for each individual cm-handle being processed (percentiles, check with  kissand )
2. Register 1 (second) batch of 100 cm-handles with no parallelism
3. Register 1 (n^th) batch of 100 cm-handles with no parallelism after several  thousand cm handles have already been registered and models synchronized
4. Repeat 3 with a parallelism of 2 or 3 to see itis effect(iveness)

Tooling, need to be decided, should be non-invasive as much as possible. Options

1. Enable logging on public method exit and entry point (@sourabh can try this)
2. Need help from ecsakoc and/or kissand to consider tools used by ESH in other use cases
   1. Visual VM  (thread focus)
   2. K6 (statistics on rest endpoints)

A/C

• Mocked number of Modules = 180
  Update dmi-plugin-perf-stub/files/moduleResourcesResponse.json to include 180 modules
• Clear reporting on each use-case reviewed by Team and stakeholder
• Tooling: Java Visual VM  (can be used locally with IDE) 
  • Report : Time spent in methods (Suggested monitor 1 batch at 100, 5000, 10000, parallelism 1 and 10)  
  • No code changes to reporting, only external tool

Out-of-Sope

1. Configuration (DB) changes, need to be handle separate