Export span links through native transport

[lloeki]

AI-generated code disclosure: this PR was implemented with substantial AI assistance and manually reviewed and validated.

What does this PR do?

Exports canonical SpanLink#to_hash payloads through the native transport, including 128-bit IDs, flags, tracestate, flattened attributes, and link order.

The conversion snapshots links and their string data before native allocation so Ruby mutation and default-proc callbacks cannot invalidate borrowed memory.

Motivation:

The native exporter does not otherwise carry span links. This adds payload parity for distributed tracing and OpenTelemetry link data.

Tracked by APMSP-3219.

Change log entry

Yes. Native trace export now supports span links.

Additional Notes:

dropped_attributes_count is intentionally deferred by additive commit 1893676d. The Agent v0.4 and indexed V1 schemas do not define it, and the existing Ruby field does not produce a non-zero value. Reverting that commit restores the native wiring if a protocol field is introduced.

Current stack position: R5 / span links, based on the common v39 uptake in #6143. M1 and events are sibling consumers; M2 follows M1, and M3 follows M2.

Stage	libdatadog API	dd-trace-rb consumer
Runtime and panic containment	#2302	#6143
M1 opaque meta_struct blobs	#2303	#6130
Span links	#2305	#6129
Span events	#2301	#6134
M2 native structured values	#2304	#6132
M3 direct structured setter	#2300	#6133

Independent fork and exception-safety prerequisites are #6127 and #6128.

How to test the change?

• Native transport suite: 145 examples, 0 failures against published v39, on Ruby 2.5 and 4.0.
• Span-link and distributed propagation coverage: 85 examples, 0 failures, 1 existing pending example.
• OpenTelemetry suite: 248 examples, 0 failures.
• Standard and Steep passed.

[datadog-official]

🎉 All green!

🧪 All tests passed
❄️ No new flaky tests detected

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 90.25% (+0.00%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: adab38d | Docs | View more details | Give us feedback!}

[dd-octo-sts]

Thank you for updating Change log entry section 👏

^{Visited at: 2026-07-30 12:12:07 UTC}

[Copilot]

Pull request overview

This PR adds native-transport support for exporting span links by snapshotting canonical SpanLink#to_hash data in the C extension and wiring it into ddog_tracer_span_set_links, with accompanying spec updates to validate wire conformance, ordering, and mutation-safety.

Changes:

• Add span-link snapshotting and FFI conversion in ext/libdatadog_api/trace_exporter.c, then set links on the Rust span during native conversion.
• Update native transport warning logic to stop warning on span links (now supported) and adjust related transport specs.
• Add/extend conformance and integration specs to assert canonical link hashes, ordering, and behavior under mutation/default-proc hazards.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
spec/datadog/tracing/transport/native/transport_spec.rb	Updates warning expectations now that native transport supports span links.
spec/datadog/tracing/transport/native/tracer_span_spec.rb	Adds native conversion behavior tests for span-link normalization, UTF-8 rejection, and snapshot cleanup.
spec/datadog/tracing/transport/native/conformance_spec.rb	Adds on-the-wire conformance checks for span-link payload content and ordering.
spec/datadog/tracing/distributed/propagation_spec.rb	Extends propagation coverage to assert SpanLink#to_hash output for extracted context.
spec/datadog/opentelemetry_spec.rb	Adds explicit assertions for OpenTelemetry link translation into canonical to_hash payloads.
lib/datadog/tracing/transport/native.rb	Removes “unsupported span links” warning path and updates commentary to reflect new support.
ext/libdatadog_api/trace_exporter.c	Implements C-side snapshotting of canonical span-link hashes and exports them via libdatadog setters.

---

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

[pr-commenter]

Benchmarks

Benchmark execution time: 2026-08-10 11:41:50

Comparing candidate commit 81dd627 in PR branch lloeki/native-span-links-consumer with baseline commit c674c96 in branch master.

📊 Benchmarking dashboard

Found 0 performance improvements and 0 performance regressions! Performance is the same for 48 metrics, 1 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 ----------------------------------'

Unstable benchmarks

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

scenario:tracing - trace.to_digest - Continue

• unstable throughput [-1210.358op/s; +1890.023op/s] or [-4.149%; +6.478%]

[p-datadog]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Chef's kiss.

Reviewed commit: adab38d60f

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".