libdatadog update to e15985f3

[dd-octo-sts]

Summary

Automated update of the libdatadog submodule to the latest HEAD.

	SHA
Previous	382a08732c4f0061c55f890830b5206afc3e929f
New	e15985f3bbcf25c29a29e4527c266b1b10942c0a

Full CI result: ❌ 15 job(s) failed
CI pipeline: https://gitlab.ddbuild.io/DataDog/apm-reliability/dd-trace-php/-/pipelines/118037424

---

libdatadog Integration Report

libdatadog SHA: e15985f3bbcf25c29a29e4527c266b1b10942c0a
Analysis date: 2026-06-11

Overall status

⚠️ Adapted (API changes fixed)

Build & test summary

15 jobs were reported as failed across the tracer and package sub-pipelines. They fall into three groups:

1. Rust compilation failures (12 jobs) — every compile tracing extension asan job (8.0–8.5, amd64 + arm64) and the compile extension: debug-zts-asan: [7.4, arm64] job failed with the same compile error in the datadog-php crate (components-rs/bytes.rs). This is a genuine libdatadog API change and is fixed below.

2. aggregate tested versions (1 job) — this aggregation step is downstream of the compile jobs; it failed only because its inputs (the compile jobs) failed. It will recover once compilation is fixed. No code change needed.

3. windows test_c: [7.2] and [7.3] (2 jobs) — both failed with job_execution_timeout after exactly 3600 s (the 1-hour job limit), not with a build/compile error. This is CI infrastructure slowness, unrelated to the libdatadog API. Treated as flaky/infra (see below).

After the fix, the datadog-php crate (the only crate in our tree that failed) should compile. rustc reports all errors in a crate before aborting and reported exactly 4, all in bytes.rs, so no other call sites in our code are affected.

Non-trivial changes made

components-rs/bytes.rs — VecMap::remove → VecMap::remove_slow

libdatadog commit 74284cac7 refactor(span)!: use VecMap for meta, metrics and meta_struct for v04 spans (#2043) changed the backing type of SpanBytes::meta, metrics and meta_struct from a hash map to the new VecMap (libdd-trace-utils/src/span/vec_map.rs).

VecMap does not expose a remove method. Removal on a vector-backed map is intentionally a linear Vec::retain operation, so the API names it remove_slow to make the cost explicit:

/// Remove all entries matching this key from the map. This method uses [Vec::retain], and is
/// thus potentially costly (like any removal in a vector-like datastructure).
pub fn remove_slow<Q>(&mut self, key: &Q)
where K: Borrow<Q>, Q: ?Sized + PartialEq { self.data.retain(|(k, _)| k.borrow() != key); }

The signature (&Q borrowed key) and semantics (remove all entries matching the key) match the old remove(&key) usage exactly, so this is a pure rename — the compiler itself suggested remove_slow. Updated the 4 call sites:

• line 249 — ddog_del_span_meta_zstr: ptr.meta.remove(...) → ptr.meta.remove_slow(...)
• line 254 — ddog_del_span_meta_str: ptr.meta.remove(...) → ptr.meta.remove_slow(...)
• line 293 — ddog_del_span_metrics_zstr: ptr.metrics.remove(...) → ptr.metrics.remove_slow(...)
• line 319 — ddog_del_span_metrics_str: ptr.metrics.remove(...) → ptr.metrics.remove_slow(...)

The other VecMap methods our code uses (insert, get, contains_key) are unchanged and still present.

Identified libdatadog issues

None identified. The compile failure was an expected breaking API change (#2043), correctly flagged with a ! in its conventional-commit title, and adapted with a direct equivalent.

Flaky / ignored failures

• windows test_c: [7.2], windows test_c: [7.3] — both ended in job_execution_timeout at the 3600 s job limit rather than a script/compile error. This is the Windows test runner being slow, not a libdatadog API problem. Ignored as infra flakiness.
• aggregate tested versions — failed only as a downstream consequence of the failed compile jobs; no independent issue. Expected to pass once the compile jobs are green.

---

/cc @bwoebi

[datadog-prod-us1-6]

 

✨ Fix all issues with BitsAI

⚠️ Warnings

🚦 16 Pipeline jobs failed

DataDog/apm-reliability/dd-trace-php | ASAN test_c with multiple observers: [8.4]     

DataDog/apm-reliability/dd-trace-php | ASAN test_c with multiple observers: [8.5]     

DataDog/apm-reliability/dd-trace-php | ASAN test_c: [8.5, arm64]     

View all 16 failed jobs.

ℹ️ Info

No other issues found (see more)

🧪 All tests passed
❄️ No new flaky tests detected

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 54.12% (+0.00%)

Useful? React with 👍 / 👎

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 63825e4 | Docs | Datadog PR Page | Give us feedback!}

[pr-commenter]

Benchmarks [ tracer ]

Benchmark execution time: 2026-06-12 02:12:19

Comparing candidate commit 63825e4 in PR branch bot/libdatadog-latest with baseline commit 759417f in branch master.

Found 3 performance improvements and 0 performance regressions! Performance is the same for 191 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:SamplingRuleMatchingBench/benchRegexMatching1

• 🟩 execution_time [-135.343ns; -71.257ns] or [-8.304%; -4.372%]

scenario:TraceSerializationBench/benchSerializeTrace

• 🟩 execution_time [-51.967µs; -36.133µs] or [-10.186%; -7.082%]

scenario:TraceSerializationBench/benchSerializeTrace-opcache

• 🟩 execution_time [-46.228µs; -30.872µs] or [-10.431%; -6.966%]