chore(profiling): [4/n] lock optimization - remove 'is_internal' concept from wrappers

[vlad-scherbich]

Description

Remove the is_internal per-lock flag from _ProfiledLock wrapper, since we will now have the concept of "excluded modules" (#16922). We can replaces it with _ALWAYS_EXCLUDED_MODULES, a list of stdlib modules implementing sync and async locks, which will never be profiled. This will effectively be the same behavior: no locks backing higher-level locks (Condition, Semaphore) will be profiled; no sample double-counting; and a perf win by removing now-unnecessary bookeeping.

What changed

Before: When a stdlib primitive (e.g. threading.Semaphore) created an internal lock, _profiled_allocate_lock would still wrap it as a _ProfiledLock, and set is_internal=True. The wrapper would exit early on each call to _acquire, _release, and _flush_sample.

After: Internal stdlib locks are never wrapped at all thanks to _ALWAYS_EXCLUDED_MODULES. _profiled_allocate_lock checks the caller's __name__ against the excluded modules and returns the native lock directly on a match.

Why this is better

	Before (is_internal)	After (_ALWAYS_EXCLUDED_MODULES)
Wrapper created for internal lock?	Yes	No
Runtime overhead per acquire/release	self.is_internal check on every call	None
Sample suppression sites	_acquire, _release, _flush_sample	—
Internal lock detection	Filename comparison (realpath per lock create)	Module name set lookup (O(1), done once at patch)

Testing

• CI
• 2 new unit tests

Risks

n/a

[cit-pr-commenter-54b7da]

Codeowners resolved as

ddtrace/internal/settings/profiling.py                                  @DataDog/profiling-python

[vlad-scherbich]

@chatgpt-codex-connector please review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: abbacc1836

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[Copilot]

Pull request overview

This PR simplifies lock profiling internals by removing _ProfiledLock.is_internal and moving stdlib internal-lock suppression to allocation time via _ALWAYS_EXCLUDED_MODULES. In the profiling codebase, that shifts stdlib lock handling from “wrap then early-exit” to “don’t wrap excluded-module locks at all”.

Changes:

• Add _ALWAYS_EXCLUDED_MODULES handling in the lock allocator and remove is_internal bookkeeping from _ProfiledLock.
• Remove asyncio collector-specific internal-module-file configuration that is no longer needed.
• Update typing stubs and add/adjust tests around always-excluded modules and wrapper shape.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
ddtrace/profiling/collector/_lock.pyx	Reworks lock allocation to skip wrapping for always-excluded modules and removes internal-lock wrapper state.
ddtrace/profiling/collector/_lock.pyi	Syncs stubs with the wrapper/API changes.
ddtrace/profiling/collector/asyncio.py	Drops obsolete asyncio internal-file detection fields.
tests/profiling/collector/test_threading.py	Updates threading lock tests for the new native-internal-lock behavior and wrapper slots.
tests/profiling/test_profiling_config.py	Adds config-level regression tests for always-excluded stdlib modules.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.