fix(profiling): fall back when PY_UNWIND is not a valid local event

[vlad-scherbich]

Description

Delta vs #19269: adds a local-event fallback for PY_UNWIND. When sys.monitoring.set_local_events rejects that event, the profiling setup retries without it; the global callback remains configured.

This PR does not add the native build or change the asyncio task-tracking path.

Layer	Meaning	Which PR
Compiled	Native prerequisite	#19269
Armed	Local-event fallback	this PR
Observable	No standalone support claim	—

Testing

Risks

Additional Notes

Sibling of the other runtime compatibility fixes. changelog/no-changelog.

[cit-pr-commenter-54b7da]

Codeowners resolved as

Resolved from the full PR diff against main using the target branch CODEOWNERS file.
CODEOWNERS team requests not listed below are not required by the current file set.

ddtrace/internal/monitoring.py                                          @DataDog/apm-core-python
tests/internal/test_monitoring.py                                       @DataDog/apm-core-python

[cit-pr-commenter-54b7da]

Circular import analysis

⚠️ Existing circular imports

There are 1 circular imports that already exist on the base branch and have not been changed by this PR.

ddtrace.errortracking._handled_exceptions.bytecode_injector -> ddtrace.errortracking._handled_exceptions.callbacks -> ddtrace.errortracking._handled_exceptions.collector -> ddtrace.errortracking._handled_exceptions.bytecode_reporting -> ddtrace.errortracking._handled_exceptions.bytecode_injector

[cit-pr-commenter-54b7da]

Dependency direction analysis

⚠️ Existing dependency direction violations

There are 230 dependency direction violations that already exist on the base branch and have not been changed by this PR.

Show existing violations (showing 5 of 230 highest severity)

ddtrace.internal.tracemethods -×-> ddtrace.trace  (internal-core -> product:tracing, score=134)
ddtrace.internal.ci_visibility.recorder -×-> ddtrace.trace  (product:ci_visibility -> product:tracing, score=132)
ddtrace.llmobs._integrations.bedrock -×-> ddtrace.trace  (product:llmobs -> product:tracing, score=132)
ddtrace.internal.ci_visibility.api._base -×-> ddtrace.trace  (product:ci_visibility -> product:tracing, score=132)
ddtrace.internal.ci_visibility.filters -×-> ddtrace.trace  (product:ci_visibility -> product:tracing, score=132)

To see all violations, download the layers-base.json and layers-pr.json artifacts from this CI job and run:

uv run --script scripts/import-analysis/layers.py compare layers-base.json layers-pr.json

[datadog-datadog-prod-us1]

⚠️ Warnings

❌ Your PR has failed checks. Please review the issues below and take necessary action before merging.

🚦 1 Pipeline job failed

Changelog | Validate changelog

View more details · View in GitHub Actions

No changelog detected for PR. Remove the 'changelog/no-changelog' label to proceed with validation.

Useful? React with 👍 / 👎

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 99d3e63 | Docs | View more details | Give us feedback!}

[pr-commenter]

Benchmarks

Benchmark execution time: 2026-09-09 20:38:43

Comparing candidate commit ccb19c1 in PR branch vlad/315-profiling-py-unwind with baseline commit 93fed02 in branch main.

📊 Benchmarking dashboard

Found 0 performance improvements and 2 performance regressions! Performance is the same for 571 metrics, 10 unstable metrics, 7 known flaky benchmarks, 17 flaky benchmarks without significant changes.

Explanation

This is an A/B test comparing a candidate commit's performance against that of a baseline commit. Performance changes are noted in the tables below as:

• 🟩 = significantly better candidate vs. baseline
• 🟥 = significantly worse candidate vs. baseline

We compute a confidence interval (CI) over the relative difference of means between metrics from the candidate and baseline commits, considering the baseline as the reference.

If the CI is entirely outside the configured SIGNIFICANT_IMPACT_THRESHOLD (or the deprecated UNCONFIDENCE_THRESHOLD), the change is considered significant.

Feel free to reach out to #apm-benchmarking-platform on Slack if you have any questions.

More details about the CI and significant changes

You can imagine this CI as a range of values that is likely to contain the true difference of means between the candidate and baseline commits.

CIs of the difference of means are often centered around 0%, because often changes are not that big:

---------------------------------(------|---^--------)-------------------------------->
                              -0.6%    0%  0.3%     +1.2%
                                 |          |        |
         lower bound of the CI --'          |        |
sample mean (center of the CI) -------------'        |
         upper bound of the CI ----------------------'

As described above, a change is considered significant if the CI is entirely outside the configured SIGNIFICANT_IMPACT_THRESHOLD (or the deprecated UNCONFIDENCE_THRESHOLD).

For instance, for an execution time metric, this confidence interval indicates a significantly worse performance:

----------------------------------------|---------|---(---------^---------)---------->
                                       0%        1%  1.3%      2.2%      3.1%
                                                  |   |         |         |
       significant impact threshold --------------'   |         |         |
                      lower bound of CI --------------'         |         |
       sample mean (center of the CI) --------------------------'         |
                      upper bound of CI ----------------------------------'

scenario:iastaspects-title_aspect

• 🟥 execution_time [+62.949µs; +68.199µs] or [+23.571%; +25.537%]

scenario:samplingrules-high_match

• 🟥 execution_time [+12.276µs; +14.263µs] or [+7.797%; +9.059%]

Unstable benchmarks

These benchmarks have a confidence interval too wide to call a change; treat them as noise rather than signal.

scenario:coreapiscenario-context_with_data_listeners

• unstable execution_time [-693.400ns; +769.937ns] or [-6.293%; +6.988%]

scenario:coreapiscenario-core_dispatch_1_listener

• unstable execution_time [-32.105ns; +33.682ns] or [-5.255%; +5.514%]

scenario:coreapiscenario-core_dispatch_50_listeners

• unstable execution_time [-1646.646ns; +1631.554ns] or [-9.668%; +9.580%]

scenario:coreapiscenario-core_dispatch_exception_listeners

• unstable execution_time [-1190.426ns; +1251.171ns] or [-9.134%; +9.600%]

scenario:coreapiscenario-core_dispatch_listeners

• unstable execution_time [-325.482ns; +323.618ns] or [-8.857%; +8.807%]

scenario:coreapiscenario-core_dispatch_no_args_listeners

• unstable execution_time [-251.828ns; +252.700ns] or [-8.594%; +8.624%]

scenario:coreapiscenario-core_dispatch_with_results_1_listener

• unstable execution_time [-71.717ns; +78.175ns] or [-6.253%; +6.816%]

scenario:coreapiscenario-core_dispatch_with_results_50_listeners

• unstable execution_time [-4042.631ns; +3980.386ns] or [-9.819%; +9.668%]

scenario:coreapiscenario-core_dispatch_with_results_listeners

• unstable execution_time [-739.823ns; +805.659ns] or [-9.092%; +9.901%]

scenario:packagesupdateimporteddependencies-import_many_stdlib_cached

• unstable execution_time [-57.483µs; +53.550µs] or [-9.847%; +9.173%]

Known flaky benchmarks

These benchmarks are marked as flaky and will not trigger a failure. Modify FLAKY_BENCHMARKS_REGEX to control which benchmarks are marked as flaky.

scenario:httppropagationinject-ids_only

• 🟥 execution_time [+2.196µs; +2.398µs] or [+13.127%; +14.338%]

scenario:iastaspects-casefold_noaspect

• 🟥 execution_time [+42.887µs; +46.358µs] or [+17.922%; +19.372%]

scenario:iastaspectsospath-ospathbasename_aspect

• 🟥 execution_time [+127.021µs; +134.417µs] or [+32.425%; +34.313%]

scenario:iastaspectssplit-rsplit_aspect

• 🟥 execution_time [+17.816µs; +20.490µs] or [+12.807%; +14.729%]

scenario:span-start

• 🟥 execution_time [+1.532ms; +1.698ms] or [+10.823%; +11.996%]

scenario:telemetryaddmetric-1-count-metric-1-times

• 🟥 execution_time [+484.791ns; +534.859ns] or [+17.645%; +19.468%]

scenario:tracer-small

• 🟥 execution_time [+37.769µs; +40.506µs] or [+11.584%; +12.423%]

Known flaky benchmarks without significant changes:

• scenario:errortrackingflasksqli-baseline
• scenario:flasksimple-iast-get
• scenario:iastaspects-casefold_aspect
• scenario:iastaspects-index_aspect
• scenario:iastaspects-ljust_noaspect
• scenario:iastaspects-lower_aspect
• scenario:iastaspects-replace_aspect
• scenario:iastaspects-rstrip_aspect
• scenario:iastaspects-swapcase_aspect
• scenario:iastaspects-title_noaspect
• scenario:iastaspects-translate_aspect
• scenario:iastaspects-translate_noaspect
• scenario:iastaspects-upper_noaspect
• scenario:packagespackageforrootmodulemapping-cache_off
• scenario:packagespackageforrootmodulemapping-cache_on
• scenario:sethttpmeta-all-enabled
• scenario:telemetryaddmetric-record-100-metrics

[Copilot]

🔵 Needs a closer look

The change alters low-level sys.monitoring event arming behavior (including global event enablement) in ways that could have non-obvious runtime and performance impact and merits final human review.

Pull request overview

Adds a runtime-compatibility fallback in ddtrace.internal.monitoring for CPython 3.15 pre-rc1 builds where PY_UNWIND is rejected as a local event, ensuring unwind callbacks still fire by enabling PY_UNWIND globally only when needed.

Changes:

• Detects set_local_events(..., PY_UNWIND) rejection and falls back to enabling PY_UNWIND via set_events() while keeping other events local.
• Adjusts _on_py_unwind behavior in fallback mode to avoid returning sticky DISABLE for unregistered code objects.
• Updates tests to accept either local-only or global-fallback arming and adds an explicit regression test for the fallback path.

File summaries

File	Description
ddtrace/internal/monitoring.py	Implements the local-event probe and global-event fallback for PY_UNWIND, plus global enable/disable recomputation.
tests/internal/test_monitoring.py	Adds fixtures/helpers and new assertions/tests to cover both local and global fallback arming behavior.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[vlad-scherbich]

Revisit if becomes a problem