chore(profiling): add recursive test to benchmarking suite

[alyshawang]

New benchmarking scenario to measure the performance impact of different DD_PROFILING_STACK_V2_MAX_FRAMES values on applications with deep call stacks.

For context, DD_PROFILING_STACK_V2_MAX_FRAMES controls the maximum number of stack frames captured and sampled by the profiler. We want to test the performance overhead of different stack frame upper limits (by running a recursive function to create the stack frames) and we are planning to increase this limit from 64 to 512 frames to provide users with more complete stack traces.

The configurable parameters are

• max_depth: controls recursion depth (stack frame count)
• enable_sleep/sleep_duration: mixed CPU/idle workload patterns
• profiler_enabled: enable profiler for frame limit testing

Checklist

• PR author has checked that all the criteria below are met
• The PR description includes an overview of the change
• The PR description articulates the motivation for the change
• The change includes tests OR the PR description describes a testing strategy
• The PR description notes risks associated with the change, if any
• Newly-added code is easy to change
• The change follows the library release note guidelines
• The change includes or references documentation updates if necessary
• Backport labels are set (if applicable)

Reviewer Checklist

• Reviewer has checked that all the criteria below are met
• Title is accurate
• All changes are related to the pull request's stated goal
• Avoids breaking API changes
• Testing strategy adequately addresses listed risks
• Newly-added code is easy to change
• Release note makes sense to a user of the library
• If necessary, author has acknowledged and discussed the performance implications of this PR as reported in the benchmarks PR comment
• Backport labels are set in a manner that is consistent with the release branch maintenance policy

[github-actions]

CODEOWNERS have been resolved as:

benchmarks/recursive_computation/config.yaml                            @DataDog/apm-core-python
benchmarks/recursive_computation/scenario.py                            @DataDog/apm-core-python
.gitlab/benchmarks/microbenchmarks.yml                                  @DataDog/python-guild @DataDog/apm-core-python

[taegyunkim]

Do you mind adding a PR description or README on what this benchmark does?

[github-actions]

Bootstrap import analysis

Comparison of import times between this PR and base.

Summary

The average import time from this PR is: 272 ± 2 ms.

The average import time from base is: 273 ± 2 ms.

The import time difference between this PR and base is: -1.36 ± 0.09 ms.

Import time breakdown

The following import paths have shrunk:

ddtrace.auto 1.945 ms (0.72%)

ddtrace.bootstrap.sitecustomize 1.274 ms (0.47%)

ddtrace.bootstrap.preload 1.274 ms (0.47%)

ddtrace.internal.remoteconfig.client 0.631 ms (0.23%)

ddtrace 0.671 ms (0.25%)

ddtrace.internal._unpatched 0.021 ms (0.01%)

[pr-commenter]

Benchmarks

Benchmark execution time: 2025-06-10 21:26:33

Comparing candidate commit a6b4e5c in PR branch alysha.wang/add_recursive_benchmarking with baseline commit e7c0d74 in branch main.

Found 0 performance improvements and 2 performance regressions! Performance is the same for 559 metrics, 3 unstable metrics.

scenario:iastaspects-format_map_aspect

• 🟥 execution_time [+439.570ns; +523.888ns] or [+13.713%; +16.344%]

scenario:iastaspectsospath-ospathsplit_aspect

• 🟥 execution_time [+713.949ns; +865.727ns] or [+14.786%; +17.929%]

[taegyunkim]

https://gitlab.ddbuild.io/DataDog/apm-reliability/dd-trace-py/-/jobs/974592671

   |
71 |                 }
72 |             )
73 |             import ddtrace.profiling.auto
   |                    ^^^^^^^^^^^^^^^^^^^^^^ F401
74 |
75 |         utils.drop_traces(tracer)
   |
   = help: Remove unused import: `ddtrace.profiling.auto`

[taegyunkim]

benchmarks/recursive_computation/scenario.py:2:8: F401 `os` imported but unused
  |
1 | import math
2 | import os
  |        ^^ F401
3 | import time
4 | from typing import Callable
  |
  = help: Remove unused import: `os`

you can run hatch run lint:fmt and hatch run lint:typing before git commit ... to address these issues.

[taegyunkim]

Looks like it's taking a long time to run on the Gitlab, let's try to figure out why and tweak it to be run < 10 minutes. We'd need to remove a few variants from the config file.