DI: Bypass subclass overrides of Exception#backtrace_locations via UnboundMethod

[p-datadog]

What does this PR do?

Adds DI::EXCEPTION_BACKTRACE_LOCATIONS, an UnboundMethod captured from Exception.instance_method(:backtrace_locations) at load time. Used via bind(exception).call to get the backtrace without dispatching through the exception's method table, bypassing subclass overrides.

Switches from Exception#backtrace (Array) to Exception#backtrace_locations (ArrayThread::Backtrace::Location). format_backtrace now reads .path, .lineno, .label directly from Location objects instead of regex-parsing formatted strings.

Follows up on this review comment on #5111.

Motivation:

DI instrumentation must never invoke customer code. serialize_throwable was calling exception.backtrace through Ruby method dispatch, which would hit subclass overrides.

The UnboundMethod approach bypasses subclass method tables entirely. It does not protect against monkeypatching Exception#backtrace_locations itself before dd-trace-rb loads, but the practical threat is subclass overrides, which it handles correctly.

Change log entry

None.

Additional Notes:

Follow-up to #5111.

How to test the change?

• spec/datadog/di/ext/exception_backtrace_spec.rb — tests UnboundMethod bypass of subclass overrides
• spec/datadog/di/probe_notification_builder_spec.rb — integration test verifying stacktrace formatting with Location objects

[github-actions]

Typing analysis

Note: Ignored files are excluded from the next sections.

steep:ignore comments

This PR introduces 1 steep:ignore comment, and clears 1 steep:ignore comment.

steep:ignore comments (+1-1)

❌ Introduced:

lib/datadog/di/probe_notification_builder.rb:137

✅ Cleared:

lib/datadog/di/probe_notification_builder.rb:141

Untyped methods

This PR introduces 2 untyped methods and 7 partially typed methods, and clears 2 untyped methods and 7 partially typed methods. It increases the percentage of typed methods from 61.69% to 61.71% (+0.02%).

Untyped methods (+2-2)

❌ Introduced:

sig/datadog/di/probe_notification_builder.rbs:51
└── def tags: () -> untyped
sig/datadog/di/probe_notification_builder.rbs:53
└── def serialized_tags: () -> untyped

✅ Cleared:

sig/datadog/di/probe_notification_builder.rbs:50
└── def tags: () -> untyped
sig/datadog/di/probe_notification_builder.rbs:52
└── def serialized_tags: () -> untyped

Partially typed methods (+7-7)

❌ Introduced:

sig/datadog/di/probe_notification_builder.rbs:32
└── def build_snapshot_base: (Context context, ?evaluation_errors: Array[untyped]?, ?captures: untyped?, ?message: String?) -> Hash[Symbol,untyped]
sig/datadog/di/probe_notification_builder.rbs:34
└── def build_condition_evaluation_failed: (Context context, untyped expr, untyped exception) -> Hash[Symbol,untyped]
sig/datadog/di/probe_notification_builder.rbs:36
└── def build_status: (Probe probe, message: untyped, status: untyped, ?exception: Exception?) -> Hash[Symbol,untyped]
sig/datadog/di/probe_notification_builder.rbs:38
└── def format_caller_locations: (Array[untyped] callers) -> Array[Hash[Symbol,untyped]]
sig/datadog/di/probe_notification_builder.rbs:40
└── def evaluate_template: (untyped template, Context context) -> [String, Array[String]]
sig/datadog/di/probe_notification_builder.rbs:42
└── def tag_process_tags!: (Hash[Symbol | String, untyped] payload, Core::Configuration::Settings settings) -> void
sig/datadog/di/probe_notification_builder.rbs:46
└── def get_local_variables: (TracePoint trace_point) -> Hash[Symbol,untyped]

✅ Cleared:

sig/datadog/di/probe_notification_builder.rbs:31
└── def build_snapshot_base: (Context context, ?evaluation_errors: Array[untyped]?, ?captures: untyped?, ?message: String?) -> Hash[Symbol,untyped]
sig/datadog/di/probe_notification_builder.rbs:33
└── def build_condition_evaluation_failed: (Context context, untyped expr, untyped exception) -> Hash[Symbol,untyped]
sig/datadog/di/probe_notification_builder.rbs:35
└── def build_status: (Probe probe, message: untyped, status: untyped, ?exception: Exception?) -> Hash[Symbol,untyped]
sig/datadog/di/probe_notification_builder.rbs:37
└── def format_caller_locations: (Array[untyped] callers) -> Array[Hash[Symbol,untyped]]
sig/datadog/di/probe_notification_builder.rbs:39
└── def evaluate_template: (untyped template, Context context) -> [String, Array[String]]
sig/datadog/di/probe_notification_builder.rbs:41
└── def tag_process_tags!: (Hash[Symbol | String, untyped] payload, Core::Configuration::Settings settings) -> void
sig/datadog/di/probe_notification_builder.rbs:45
└── def get_local_variables: (TracePoint trace_point) -> Hash[Symbol,untyped]

If you believe a method or an attribute is rightfully untyped or partially typed, you can add # untyped:accept on the line before the definition to remove it from the stats.

[pr-commenter]

Benchmarks

Benchmark execution time: 2026-03-30 13:08:52

Comparing candidate commit 59fe0b5 in PR branch di-c-ext-exception-backtrace with baseline commit 9220fe8 in branch master.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 46 metrics, 0 unstable metrics.

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 ----------------------------------'

[datadog-datadog-prod-us1-2]

✅ Tests

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 95.36% (-0.00%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: e574247 | Docs | Datadog PR Page | Was this helpful? React with 👍/👎 or give us feedback!}

[p-datadog]

@codex review

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.

[p-datadog]

@copilot review

[Copilot]

Pull request overview

This PR bypasses subclass overrides of Exception#backtrace_locations by capturing an UnboundMethod from the core Exception class at load time. The change switches from parsing backtrace strings via regex to working directly with structured Thread::Backtrace::Location objects.

Changes:

• Adds DI::EXCEPTION_BACKTRACE_LOCATIONS UnboundMethod constant to bypass subclass overrides
• Updates format_backtrace to process Location objects directly instead of regex-parsing strings
• Removes the now-unused BACKTRACE_FRAME_PATTERN regex constant
• Updates type signatures and adds comprehensive test coverage

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
lib/datadog/di.rb	Adds EXCEPTION_BACKTRACE_LOCATIONS UnboundMethod constant
lib/datadog/di/probe_notification_builder.rb	Updates backtrace handling to use Location objects and UnboundMethod
sig/datadog/di.rbs	Declares EXCEPTION_BACKTRACE_LOCATIONS type
sig/datadog/di/probe_notification_builder.rbs	Updates format_backtrace parameter type
spec/datadog/di/ext/exception_backtrace_spec.rb	Tests UnboundMethod bypasses subclass overrides
spec/datadog/di/probe_notification_builder_spec.rb	Integration test for stacktrace with overridden backtrace method

---

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

[p-datadog]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 59fe0b52c0

ℹ️ About Codex in GitHub

Your team has set up Codex to 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 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".