feat(tracing): add DD_TRACE_PROPAGATION_EXTRACT_FIRST by MilanGarnier · Pull Request #356 · DataDog/dd-trace-cpp

MilanGarnier · 2026-08-19T09:41:34Z

Description

Adds C++ support for DD_TRACE_PROPAGATION_EXTRACT_FIRST.
Existing documentation for the configuration : DD_TRACE_PROPAGATION_EXTRACT_FIRST in Feature Parity Dashboard

Motivation

Allow extraction to stop after the first successfully extracted trace context, matching the shared tracer configuration.

Additional Notes

Priority: low

Generated supported-configurations.json with config-inversion.
Unit tests cover configuration precedence, heterogeneous W3C tracestate handling, and fallback after an unsuccessful extractor.

Jira ticket: N/A

pr-commenter · 2026-08-19T09:48:08Z

Benchmarks

Benchmark execution time: 2026-08-26 13:10:45

Comparing candidate commit 984a97f in PR branch milan.garnier/propagation-extract-first with baseline commit 765983d in branch main.

Found 2 performance improvements and 3 performance regressions! Performance is the same for 3 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 ----------------------------------'

scenario:BM_HexPadded_uint64/NoPadding

🟥 execution_time [+2.318ns; +2.362ns] or [+4.429%; +4.513%]

scenario:BM_HexPadded_uint64/WorstCasePadding

🟥 execution_time [+1.282ns; +1.373ns] or [+2.677%; +2.866%]

scenario:BM_Hex_uint64

🟥 execution_time [+0.487ns; +0.510ns] or [+2.119%; +2.220%]

scenario:BM_TraceID_ParseHex/128bit

🟩 execution_time [-49.662ns; -49.610ns] or [-28.753%; -28.723%]

scenario:BM_TraceID_ParseHex/64bit

🟩 execution_time [-22.507ns; -22.491ns] or [-24.526%; -24.509%]

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9f8c667bb5

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".

chatgpt-codex-connector · 2026-08-26T09:21:51Z

+    if (propagation_extract_first_ && extracted_trace_context) {
+      break;


Validate context before stopping extraction

When DD_TRACE_PROPAGATION_EXTRACT_FIRST=true, an earlier style can return a trace ID without a usable parent ID or origin—for example, incomplete Datadog headers followed by a complete matching W3C traceparent. This condition stops immediately because it checks only trace_id, after which the validation below returns MISSING_PARENT_SPAN_ID instead of allowing the later style to supply a valid context. Stop only after the extracted context meets the validity requirements used below; otherwise continue trying configured styles.

Useful? React with 👍 / 👎.

Indeed, we can see line 375 a test about parent_id.
I think we should clarify what is a successful extraction (and I suppose this would deserve a dedicated unit test).
Once again, I would really like to clarify the purpose of this block of code. I would start by extracting lines 277-343, but that's only a suggestion after a quick analysis, to be challenged.

xlamorlette-datadog · 2026-08-26T09:57:53Z

For the record: DD_TRACE_PROPAGATION_EXTRACT_FIRST in Feature Parity Dashboard.

xlamorlette-datadog

I find the C++ Tracer code difficult to reason because it is not in a very good state: it contains lot of legacy code, with questionable technical choices (but, unfortunately, many are part of the public interface).
From now on, I think we should try to clean the existing code to avoid deteriorating the situation further.
Especially here: Tracer::extract_span() is the biggest function: 273 lines!
Please, I think we should refactor this function (in a dedicated PR) before adding complexity to it.

xlamorlette-datadog · 2026-08-26T10:35:18Z

                                  {extracted_tag});

-    if (!first_style_with_trace_id && data->trace_id.has_value()) {
+    const bool extracted_trace_context = data->trace_id.has_value();


I think the extracted_trace_context name is misleading:

It makes me think this variable is a trace context.

This only checks that a trace id was extracted, not a full context, which, I think, relates to the comment from Codex below.

Actually, because the existing code is not clean, I find it very difficult to reason about this:

The data name does not convey any useful meaning.

The on-going modification is only about a part of the huge (270 lines!) Tracer::extract_span() function.

I suggest adding a dedicated refactoring PR before this one to clean the Tracer::extract_span() function:

At least extract the part to be modified, to clarify its purpose, and improve variable names. This could also allow to unit test easier the new behavior.

Ideally, split the function in big blocks (skimming the function, it seems to me there are at least 4 different blocks).

xlamorlette-datadog · 2026-08-26T12:13:06Z

+    if (propagation_extract_first_ && extracted_trace_context) {
+      break;


Indeed, we can see line 375 a test about parent_id.
I think we should clarify what is a successful extraction (and I suppose this would deserve a dedicated unit test).
Once again, I would really like to clarify the purpose of this block of code. I would start by extracting lines 277-343, but that's only a suggestion after a quick analysis, to be challenged.

xlamorlette-datadog · 2026-08-26T12:15:11Z

    std::vector<PropagationStyle> injection_styles;
    std::unordered_map<std::string, std::string> extracted_headers;
    std::unordered_map<std::string, std::string> expected_injected_headers;
+    bool extract_first = false;


This should be added in the capture block (lines 2121-2126).

feat(tracing): add DD_TRACE_PROPAGATION_EXTRACT_FIRST

9f8c667

MilanGarnier temporarily deployed to dev August 19, 2026 09:42 — with GitHub Actions Inactive

MilanGarnier temporarily deployed to dev August 19, 2026 09:46 — with GitHub Actions Inactive

This comment has been minimized.

Sign in to view

MilanGarnier added AI Generated Largely based on code generated by an AI or LLM. This label is the same across all dd-trace-* repos and removed AI Generated Largely based on code generated by an AI or LLM. This label is the same across all dd-trace-* repos labels Aug 19, 2026

MilanGarnier marked this pull request as ready for review August 26, 2026 09:16

MilanGarnier requested review from a team as code owners August 26, 2026 09:16

MilanGarnier requested a review from cataphract August 26, 2026 09:16

chatgpt-codex-connector Bot reviewed Aug 26, 2026

View reviewed changes

xlamorlette-datadog requested changes Aug 26, 2026

View reviewed changes

fix(propagation): validate extract-first contexts

984a97f

MilanGarnier marked this pull request as draft August 26, 2026 13:23

Navigation Menu

Uh oh!

feat(tracing): add DD_TRACE_PROPAGATION_EXTRACT_FIRST - #356

feat(tracing): add DD_TRACE_PROPAGATION_EXTRACT_FIRST#356
MilanGarnier wants to merge 2 commits into
mainfrom
milan.garnier/propagation-extract-first

MilanGarnier commented Aug 19, 2026 •

edited

Loading

Uh oh!

pr-commenter Bot commented Aug 19, 2026 •

edited

Loading

Explanation

More details about the CI and significant changes

Uh oh!

This comment has been minimized.

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Aug 26, 2026

Uh oh!

xlamorlette-datadog Aug 26, 2026

Uh oh!

xlamorlette-datadog commented Aug 26, 2026

Uh oh!

xlamorlette-datadog left a comment

Uh oh!

xlamorlette-datadog Aug 26, 2026

Uh oh!

xlamorlette-datadog Aug 26, 2026

Uh oh!

xlamorlette-datadog Aug 26, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		if (propagation_extract_first_ && extracted_trace_context) {
		break;

Uh oh!

Conversation

MilanGarnier commented Aug 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Motivation

Additional Notes

Uh oh!

pr-commenter Bot commented Aug 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Benchmarks

Explanation

More details about the CI and significant changes

scenario:BM_HexPadded_uint64/NoPadding

scenario:BM_HexPadded_uint64/WorstCasePadding

scenario:BM_Hex_uint64

scenario:BM_TraceID_ParseHex/128bit

scenario:BM_TraceID_ParseHex/64bit

Uh oh!

This comment has been minimized.

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Aug 26, 2026

Choose a reason for hiding this comment

Uh oh!

xlamorlette-datadog Aug 26, 2026

Choose a reason for hiding this comment

Uh oh!

xlamorlette-datadog commented Aug 26, 2026

Uh oh!

xlamorlette-datadog left a comment

Choose a reason for hiding this comment

Uh oh!

xlamorlette-datadog Aug 26, 2026

Choose a reason for hiding this comment

Uh oh!

xlamorlette-datadog Aug 26, 2026

Choose a reason for hiding this comment

Uh oh!

xlamorlette-datadog Aug 26, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

MilanGarnier commented Aug 19, 2026 •

edited

Loading

pr-commenter Bot commented Aug 19, 2026 •

edited

Loading