[tools] Stress test improvements

[CodeBlanch]

Some improvements to the stress tests I made to help in testing #5364.

Changes

• There is now a CLI which allows you to do this type of thing: OpenTelemetry.Tests.Stress.Metrics.exe -t Counter -v -o -e -d 300
• While the test is running you will now get some statistics on the console by default.
• Introduced some class hierarchy instead of linking in the Skeleton/Program from the demo test into the actual tests.

Options

Shared by all signals:

  -c, --concurrency      The concurrency (number of executing threads) for the stress test. Default value: Environment.ProcessorCount.

  -p, --internal_port    The prometheus http listener port where prometheus will be exposed for retrieving internal metrics while the stress test is running. Set to '0' to disable. Default value: 9464.

  -d, --duration         The duration for the stress test to run in seconds. If set to '0' or a negative value the stress test will run until canceled. Default value: 0.

Metrics:

  -t, --type             The metrics stress test type to run. Valid values: [Histogram, Counter]. Default value: Histogram.

  -m, --metrics_port     The prometheus http listener port where prometheus will be exposed for retrieving test metrics while the stress test is running. Set to '0' to disable. Default value: 9185.

  -v, --view             Whether or not a view should be configured to filter tags for the stress test. Default value: False.

  -o, --otlp             Whether or not an OTLP exporter should be added for the stress test. Default value: False.

  -i, --interval         The OTLP exporter export interval in milliseconds. Default value: 5000.

  -e, --exemplars        Whether or not to enable exemplars for the stress test. Default value: False.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 83.22%. Comparing base (6250307) to head (93953fd).
Report is 97 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5381      +/-   ##
==========================================
- Coverage   83.38%   83.22%   -0.16%     
==========================================
  Files         297      279      -18     
  Lines       12531    11961     -570     
==========================================
- Hits        10449     9955     -494     
+ Misses       2082     2006      -76     

Flag	Coverage Δ
unittests	?
unittests-Solution-Experimental	83.22% <ø> (?)
unittests-Solution-Stable	82.97% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

see 48 files with indirect coverage changes

[reyang]

Update the README file as well.

[CodeBlanch]

Forgot to ask, was "Runaway" intentional? In the running output we have "Runway" so it seemed like a typo.

[reyang]

I don't know what's the perfect term, I just picked a term that seems to fit the intention http://osr507doc.xinuos.com/en/HANDBOOK/runaway_proc.html. (e.g. if user keeps this process, it'll consume lots of system resources)

The comment here is more emphasizing on "whatever term we choose, we have to be consistent".

[CodeBlanch]

I just switched it to "Running Time" in both spots. Seemed less likely to cause any confusion but I don't absolutely love it 😄

Let me do README updates in a follow-up as this is getting a bit big already?

[utpilla]

Forgot to ask, was "Runaway" intentional? In the running output we have "Runway" so it seemed like a typo.

#3496 (comment) 😀

[rajkumar-rangaraj]

LGTM