[PROF-10589] Add microbenchmark for GVL profiling sampling

[ivoanjo]

What does this PR do?

This PR adds a new profiler microbenchmark that tests the performance of sampling as triggered from the GVL profiling feature.

Motivation:

I've created this benchmark as part of validating the performance of this feature.

Additional Notes:

Profiling the profiler with ddprof, it looks like right now the biggest performance bottleneck is the cost of stack sampling.

This can be validated easily -- changing the depth of the stack for the testing thread makes a big difference.

How to test the change?

Validate that this benchmark gets picked up by our usual harness to run microbenchmarks.

[datadog-datadog-prod-us1]

⚪ Code Quality Violation

Suggested change

	$stderr.puts "Skipping benchmark because it requires Ruby 3.2 or newer"
	warn "Skipping benchmark because it requires Ruby 3.2 or newer"

Do not print on stderr directly (...read more)

In Ruby, it is a good practice to use warn instead of $stderr.puts for issuing warning messages. The warn method is specifically designed for this purpose and its use makes the intention of the code clearer.

The importance of this rule lies in the fact that warn and $stderr.puts behave differently in certain situations. For example, warn will prepend the filename and line number to the warning message, which can be very helpful for debugging. Moreover, warn respects the -W command-line option for setting warning levels, while $stderr.puts does not.

To avoid violating this rule, replace any instances of $stderr.puts with warn when you want to issue a warning. Remember that the purpose of $stderr.puts is to write to the standard error, not to issue warnings. If you want to write to the standard error for reasons other than issuing warnings, $stderr.puts is the appropriate method to use.

    

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.85%. Comparing base (58b17f0) to head (5585532).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3989      +/-   ##
==========================================
- Coverage   97.87%   97.85%   -0.02%     
==========================================
  Files        1314     1314              
  Lines       78652    78652              
  Branches     3909     3909              
==========================================
- Hits        76978    76966      -12     
- Misses       1674     1686      +12     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pr-commenter]

Benchmarks

Benchmark execution time: 2024-10-10 15:32:52

Comparing candidate commit 5585532 in PR branch ivoanjo/prof-10589-gvl-profiling-microbenchmark with baseline commit 58b17f0 in branch master.

Found 0 performance improvements and 1 performance regressions! Performance is the same for 23 metrics, 2 unstable metrics.

scenario:profiler - sample timeline=false

• 🟥 throughput [-0.430op/s; -0.398op/s] or [-6.246%; -5.773%]