feat(profiling)!: expose missing API in CXX by danielsn · Pull Request #2497 · DataDog/libdatadog

danielsn · 2026-09-09T19:14:57Z

What does this PR do?

Updates the CXX API to expose APIs for

dictionary-backed samples
timestamped samples
split profile uploading

Motivation

I would like to use this API for Python, but currently we're missing this functionality so we have to use the old C api.

Additional Notes

Anything else we should know when reviewing?

How to test the change?

Describe here in detail how the change can be validated.

github-actions · 2026-09-09T19:17:35Z

📚 Documentation Check Results

⚠️ 1419 documentation warning(s) found

📦 `libdd-profiling-ffi` - 770 warning(s)

📦 `libdd-profiling` - 649 warning(s)

Updated: 2026-09-10 20:21:10 UTC | Commit: 3f0efda | missing-docs job results

danielsn · 2026-09-09T19:17:21Z

+        // NOTE: api2's Rust/C FFI datatype uses file_name, while pprof and
+        // the string-based CXX Function use filename. Keep file_name for now
+        // to match api2; revisit the CXX-facing name during PR review.
+        file_name: StringId2,


Anyone have any thoughts here?

github-actions · 2026-09-09T19:18:27Z

🔒 Cargo Deny Results

⚠️ 8 issue(s) found, showing only errors (advisories, bans, sources)

📦 `libdd-profiling-ffi` - 4 error(s)

Show output

error[unsound]: `Bitmap::try_from(&[u8])` can create invalid values
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:20:1
   │
20 │ bitmaps 3.2.1 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ unsound advisory detected
   │
   ├ ID: RUSTSEC-2025-0167
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2025-0167
   ├ The `TryFrom<&[u8]>` implementation for `Bitmap<SIZE>` copies the input bytes
     into an uninitialized backing store and calls `assume_init()` without
     validating that the bytes form a valid value of the backing store type. For
     `SIZE = 1` the backing store is a `bool`, so any input byte other than `0x00`
     or `0x01` produces an invalid value, which is immediate undefined behavior.
     
     The `AsMut<[u8]>` implementation has the same problem, as it allows safe code
     to write invalid bit patterns into the backing store through the returned slice.
     
     No fixed version is available, as the crate is unmaintained; its GitHub
     repository was archived by the owner on 2026-05-03.
   ├ Announcement: https://github.com/bodil/bitmaps/issues/35
   ├ Solution: No safe upgrade is available!
   ├ bitmaps v3.2.1
     └── libdd-profiling v1.0.0
         ├── (dev) libdd-profiling v1.0.0 (*)
         └── libdd-profiling-ffi v1.0.0

error[unmaintained]: bitmaps is unmaintained
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:20:1
   │
20 │ bitmaps 3.2.1 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ unmaintained advisory detected
   │
   ├ ID: RUSTSEC-2026-0247
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0247
   ├ The bitmaps crate is unmaintained; all versions are affected. The GitHub
     repository was archived by the owner on 2026-05-03.
     
     Recommended alternatives:
     
     * [fixedbitset](https://crates.io/crates/fixedbitset)
     * [bitvec](https://crates.io/crates/bitvec)
   ├ Announcement: https://github.com/bodil/bitmaps
   ├ Solution: No safe upgrade is available!
   ├ bitmaps v3.2.1
     └── libdd-profiling v1.0.0
         ├── (dev) libdd-profiling v1.0.0 (*)
         └── libdd-profiling-ffi v1.0.0

error[vulnerability]: NSEC3 closest-encloser proof validation enters unbounded loop on cross-zone responses
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:91:1
   │
91 │ hickory-proto 0.25.2 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ security vulnerability detected
   │
   ├ ID: RUSTSEC-2026-0118
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0118
   ├ The NSEC3 closest-encloser proof validation in `hickory-proto`'s
     `DnssecDnsHandle` walks from the QNAME up to the SOA owner name, building a
     list of candidate encloser names. The iterator used assumes the
     QNAME is a descendant of the SOA owner, terminating only when the current
     candidate equals the SOA name. When the SOA in a response's authority section
     is not an ancestor of the QNAME, the loop stalls at the DNS root and never
     terminates, repeatedly calling `Name::base_name()` and pushing newly allocated
     `Name` and hashed-name entries into the candidate `Vec`.
     
     The bug is reachable by any caller of `DnssecDnsHandle` — including the
     resolver, recursor, and client — when built with the `dnssec-ring` or
     `dnssec-aws-lc-rs` feature and configured to perform DNSSEC validation. It is
     triggered while validating a NoData or NXDomain response whose authority
     section contains an SOA record from a zone other than an ancestor of the
     QNAME, on a code path that requires NSEC3 closest-encloser proof. In practice
     this can be reached through an insecure CNAME chain that crosses zone
     boundaries into a DNSSEC-signed zone returning NoData, but the minimum
     condition is just a mismatched SOA owner on a response requiring NSEC3
     validation.
     
     A `debug_assert_ne!(name, Name::root())` guards the loop body, so debug builds
     abort with a panic on the first iteration past the root. Release builds
     compile the assertion out and run the loop unbounded, allocating until the
     process exhausts available memory (OOM). A reachable upstream attacker who
     can return such a response can therefore crash a debug-built validator or
     exhaust memory on a release-built one.
     
     The affected code was migrated from `hickory-proto` to `hickory-net` as part of
     the 0.26.0 release. The `hickory-proto` 0.26.x release no longer offers
     `DnssecDnsHandle` and so we recommend all affected users update to `hickory-net`
     0.26.1 when the implementation of that type is required.
   ├ Announcement: https://github.com/hickory-dns/hickory-dns/security/advisories/GHSA-3v94-mw7p-v465
   ├ Solution: No safe upgrade is available!
   ├ hickory-proto v0.25.2
     └── hickory-resolver v0.25.2
         └── reqwest v0.13.2
             ├── libdd-common v6.0.0
             │   ├── libdd-common-ffi v44.0.0
             │   │   └── libdd-profiling-ffi v1.0.0
             │   ├── libdd-profiling v1.0.0
             │   │   ├── (dev) libdd-profiling v1.0.0 (*)
             │   │   └── libdd-profiling-ffi v1.0.0 (*)
             │   └── libdd-profiling-ffi v1.0.0 (*)
             └── libdd-profiling v1.0.0 (*)

error[vulnerability]: CPU exhaustion during message encoding due to O(n²) name compression
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:91:1
   │
91 │ hickory-proto 0.25.2 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ security vulnerability detected
   │
   ├ ID: RUSTSEC-2026-0119
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0119
   ├ During message encoding, `hickory-proto`'s `BinEncoder` stores pointers to
     labels that are candidates for name compression in a `Vec<(usize, Vec<u8>)>`.
     The name compression logic then searches for matches with a linear scan.
     
     A malicious message with many records can both introduce many candidate labels,
     and invoke this linear scan many times. This can amplify CPU exhaustion in DoS
     attacks.
     
     This is similar to
     [CVE-2024-8508](https://www.nlnetlabs.nl/downloads/unbound/CVE-2024-8508.txt).
     
     We recommend all affected users update to `hickory-proto` 0.26.1 for the fix.
   ├ Announcement: https://github.com/hickory-dns/hickory-dns/security/advisories/GHSA-q2qq-hmj6-3wpp
   ├ Solution: Upgrade to >=0.26.1 (try `cargo update -p hickory-proto`)
   ├ hickory-proto v0.25.2
     └── hickory-resolver v0.25.2
         └── reqwest v0.13.2
             ├── libdd-common v6.0.0
             │   ├── libdd-common-ffi v44.0.0
             │   │   └── libdd-profiling-ffi v1.0.0
             │   ├── libdd-profiling v1.0.0
             │   │   ├── (dev) libdd-profiling v1.0.0 (*)
             │   │   └── libdd-profiling-ffi v1.0.0 (*)
             │   └── libdd-profiling-ffi v1.0.0 (*)
             └── libdd-profiling v1.0.0 (*)

advisories FAILED, bans ok, sources ok

📦 `libdd-profiling` - 4 error(s)

Show output

error[unsound]: `Bitmap::try_from(&[u8])` can create invalid values
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:15:1
   │
15 │ bitmaps 3.2.1 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ unsound advisory detected
   │
   ├ ID: RUSTSEC-2025-0167
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2025-0167
   ├ The `TryFrom<&[u8]>` implementation for `Bitmap<SIZE>` copies the input bytes
     into an uninitialized backing store and calls `assume_init()` without
     validating that the bytes form a valid value of the backing store type. For
     `SIZE = 1` the backing store is a `bool`, so any input byte other than `0x00`
     or `0x01` produces an invalid value, which is immediate undefined behavior.
     
     The `AsMut<[u8]>` implementation has the same problem, as it allows safe code
     to write invalid bit patterns into the backing store through the returned slice.
     
     No fixed version is available, as the crate is unmaintained; its GitHub
     repository was archived by the owner on 2026-05-03.
   ├ Announcement: https://github.com/bodil/bitmaps/issues/35
   ├ Solution: No safe upgrade is available!
   ├ bitmaps v3.2.1
     └── libdd-profiling v1.0.0
         └── (dev) libdd-profiling v1.0.0 (*)

error[unmaintained]: bitmaps is unmaintained
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:15:1
   │
15 │ bitmaps 3.2.1 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ unmaintained advisory detected
   │
   ├ ID: RUSTSEC-2026-0247
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0247
   ├ The bitmaps crate is unmaintained; all versions are affected. The GitHub
     repository was archived by the owner on 2026-05-03.
     
     Recommended alternatives:
     
     * [fixedbitset](https://crates.io/crates/fixedbitset)
     * [bitvec](https://crates.io/crates/bitvec)
   ├ Announcement: https://github.com/bodil/bitmaps
   ├ Solution: No safe upgrade is available!
   ├ bitmaps v3.2.1
     └── libdd-profiling v1.0.0
         └── (dev) libdd-profiling v1.0.0 (*)

error[vulnerability]: NSEC3 closest-encloser proof validation enters unbounded loop on cross-zone responses
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:80:1
   │
80 │ hickory-proto 0.25.2 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ security vulnerability detected
   │
   ├ ID: RUSTSEC-2026-0118
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0118
   ├ The NSEC3 closest-encloser proof validation in `hickory-proto`'s
     `DnssecDnsHandle` walks from the QNAME up to the SOA owner name, building a
     list of candidate encloser names. The iterator used assumes the
     QNAME is a descendant of the SOA owner, terminating only when the current
     candidate equals the SOA name. When the SOA in a response's authority section
     is not an ancestor of the QNAME, the loop stalls at the DNS root and never
     terminates, repeatedly calling `Name::base_name()` and pushing newly allocated
     `Name` and hashed-name entries into the candidate `Vec`.
     
     The bug is reachable by any caller of `DnssecDnsHandle` — including the
     resolver, recursor, and client — when built with the `dnssec-ring` or
     `dnssec-aws-lc-rs` feature and configured to perform DNSSEC validation. It is
     triggered while validating a NoData or NXDomain response whose authority
     section contains an SOA record from a zone other than an ancestor of the
     QNAME, on a code path that requires NSEC3 closest-encloser proof. In practice
     this can be reached through an insecure CNAME chain that crosses zone
     boundaries into a DNSSEC-signed zone returning NoData, but the minimum
     condition is just a mismatched SOA owner on a response requiring NSEC3
     validation.
     
     A `debug_assert_ne!(name, Name::root())` guards the loop body, so debug builds
     abort with a panic on the first iteration past the root. Release builds
     compile the assertion out and run the loop unbounded, allocating until the
     process exhausts available memory (OOM). A reachable upstream attacker who
     can return such a response can therefore crash a debug-built validator or
     exhaust memory on a release-built one.
     
     The affected code was migrated from `hickory-proto` to `hickory-net` as part of
     the 0.26.0 release. The `hickory-proto` 0.26.x release no longer offers
     `DnssecDnsHandle` and so we recommend all affected users update to `hickory-net`
     0.26.1 when the implementation of that type is required.
   ├ Announcement: https://github.com/hickory-dns/hickory-dns/security/advisories/GHSA-3v94-mw7p-v465
   ├ Solution: No safe upgrade is available!
   ├ hickory-proto v0.25.2
     └── hickory-resolver v0.25.2
         └── reqwest v0.13.2
             ├── libdd-common v6.0.0
             │   └── libdd-profiling v1.0.0
             │       └── (dev) libdd-profiling v1.0.0 (*)
             └── libdd-profiling v1.0.0 (*)

error[vulnerability]: CPU exhaustion during message encoding due to O(n²) name compression
   ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:80:1
   │
80 │ hickory-proto 0.25.2 registry+https://github.com/rust-lang/crates.io-index
   │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ security vulnerability detected
   │
   ├ ID: RUSTSEC-2026-0119
   ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0119
   ├ During message encoding, `hickory-proto`'s `BinEncoder` stores pointers to
     labels that are candidates for name compression in a `Vec<(usize, Vec<u8>)>`.
     The name compression logic then searches for matches with a linear scan.
     
     A malicious message with many records can both introduce many candidate labels,
     and invoke this linear scan many times. This can amplify CPU exhaustion in DoS
     attacks.
     
     This is similar to
     [CVE-2024-8508](https://www.nlnetlabs.nl/downloads/unbound/CVE-2024-8508.txt).
     
     We recommend all affected users update to `hickory-proto` 0.26.1 for the fix.
   ├ Announcement: https://github.com/hickory-dns/hickory-dns/security/advisories/GHSA-q2qq-hmj6-3wpp
   ├ Solution: Upgrade to >=0.26.1 (try `cargo update -p hickory-proto`)
   ├ hickory-proto v0.25.2
     └── hickory-resolver v0.25.2
         └── reqwest v0.13.2
             ├── libdd-common v6.0.0
             │   └── libdd-profiling v1.0.0
             │       └── (dev) libdd-profiling v1.0.0 (*)
             └── libdd-profiling v1.0.0 (*)

advisories FAILED, bans ok, sources ok

Updated: 2026-09-10 20:22:38 UTC | Commit: 3f0efda | dependency-check job results

datadog-datadog-prod-us1 · 2026-09-09T19:20:40Z

✅ All CI checks and tests passed.

🎉 All green!

🧪 All tests passed
❄️ No new flaky tests detected

🎯 Code Coverage (details)
• Patch Coverage: 96.66%
• Overall Coverage: 78.02% (+0.38%)

_{This comment will be updated automatically if new data arrives.

🔗 Commit SHA: bb7ab76 | Docs | View more details | Give us feedback!}

morrisonlevi · 2026-09-09T19:24:26Z

                // SAFETY: all ids come from the profile's dictionary.
                black_box(unsafe {
-                    profile.try_add_sample2(&locations, &values, labels_iter, None)
+                    profile.try_add_sample2(locations.iter().copied(), &values, labels_iter, None)


Why did this have to change, by the way? I assume it's because &[Location] doesn't implement an iterator, but rather IntoIterator?

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7aeac76d76

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 · 2026-09-09T19:25:47Z

+        fn create_with_dictionary(
+            sample_types: Vec<SampleType>,
+            period: &Period,
+            dictionary: &ProfilesDictionary,
+        ) -> Result<Box<Profile>>;


Provide a no-period dictionary profile constructor

Dictionary-backed CXX profiles now require a Period, so consumers of the existing Profile::create_no_period path cannot migrate to add_sample2; supplying a dummy period changes the serialized pprof by emitting period fields. The internal try_new_with_dictionary and C FFI dictionary constructor both support None, so expose an equivalent dictionary-backed no-period factory here.

Useful? React with 👍 / 👎.

pr-commenter · 2026-09-09T19:28:59Z

Benchmarks

Comparison

Benchmark execution time: 2026-09-10 20:28:56

Comparing candidate commit bb7ab76 in PR branch better-cxx with baseline commit 8b0fa65 in branch main.

📊 Benchmarking dashboard

Found 0 performance improvements and 0 performance regressions! Performance is the same for 13 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 ----------------------------------'

Candidate

Candidate benchmark details

Group 1

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`bb7ab76`	1789071525	better-cxx

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profiles_dictionary/profile_string_inserts/threads/1	execution_time	374.306µs	375.407µs ± 0.877µs	375.219µs ± 0.604µs	375.944µs	376.755µs	376.815µs	376.830µs	0.43%	0.430	-1.019	0.22%	0.277µs	1	10
profiles_dictionary/profile_string_inserts/threads/1	throughput	2717402.651op/s	2727717.006op/s ± 6366.589op/s	2729070.565op/s ± 4403.500op/s	2731513.405op/s	2735475.402op/s	2735681.362op/s	2735732.852op/s	0.24%	-0.424	-1.021	0.22%	2013.292op/s	1	10
profiles_dictionary/profile_string_inserts/threads/16	execution_time	1.989ms	2.001ms ± 0.005ms	2.002ms ± 0.003ms	2.003ms	2.006ms	2.007ms	2.007ms	0.27%	-1.143	1.084	0.24%	0.002ms	1	10
profiles_dictionary/profile_string_inserts/threads/16	throughput	8162765.196op/s	8189930.994op/s ± 21162.055op/s	8184982.677op/s ± 12194.106op/s	8198383.151op/s	8221844.454op/s	8235839.831op/s	8239338.675op/s	0.66%	1.156	1.109	0.25%	6692.029op/s	1	10
profiles_dictionary/profile_string_inserts/threads/2	execution_time	635.919µs	639.286µs ± 3.136µs	638.742µs ± 1.429µs	639.729µs	644.725µs	644.976µs	645.038µs	0.99%	0.888	-0.434	0.47%	0.992µs	1	10
profiles_dictionary/profile_string_inserts/threads/2	throughput	3175005.395op/s	3203641.956op/s ± 15652.119op/s	3206300.735op/s ± 7180.489op/s	3214092.297op/s	3220093.251op/s	3220447.671op/s	3220536.276op/s	0.44%	-0.877	-0.445	0.46%	4949.635op/s	1	10
profiles_dictionary/profile_string_inserts/threads/4	execution_time	814.065µs	817.919µs ± 3.567µs	817.132µs ± 2.652µs	820.490µs	823.165µs	823.460µs	823.533µs	0.78%	0.395	-1.355	0.41%	1.128µs	1	10
profiles_dictionary/profile_string_inserts/threads/4	throughput	4973691.562op/s	5007917.371op/s ± 21803.487op/s	5012655.987op/s ± 16268.677op/s	5027812.025op/s	5030248.012op/s	5031282.282op/s	5031540.849op/s	0.38%	-0.389	-1.362	0.41%	6894.868op/s	1	10

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profiles_dictionary/profile_string_inserts/threads/1	execution_time	[374.864µs; 375.951µs] or [-0.145%; +0.145%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/1	throughput	[2723771.025op/s; 2731662.986op/s] or [-0.145%; +0.145%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/16	execution_time	[1.997ms; 2.004ms] or [-0.160%; +0.160%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/16	throughput	[8176814.857op/s; 8203047.130op/s] or [-0.160%; +0.160%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/2	execution_time	[637.342µs; 641.230µs] or [-0.304%; +0.304%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/2	throughput	[3193940.850op/s; 3213343.061op/s] or [-0.303%; +0.303%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/4	execution_time	[815.708µs; 820.130µs] or [-0.270%; +0.270%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/4	throughput	[4994403.678op/s; 5021431.064op/s] or [-0.270%; +0.270%]	None	None	None

Group 2

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`bb7ab76`	1789071525	better-cxx

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_add_sample_timestamped_x1000	execution_time	4.050ms	4.056ms ± 0.009ms	4.055ms ± 0.001ms	4.056ms	4.059ms	4.062ms	4.176ms	3.00%	12.076	157.210	0.22%	0.001ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_add_sample_timestamped_x1000	execution_time	[4.054ms; 4.057ms] or [-0.031%; +0.031%]	None	None	None

Group 3

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`bb7ab76`	1789071525	better-cxx

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
benching string interning on wordpress profile	execution_time	123.703µs	124.813µs ± 0.468µs	124.741µs ± 0.141µs	124.898µs	125.374µs	126.467µs	128.915µs	3.35%	4.147	31.231	0.37%	0.033µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
benching string interning on wordpress profile	execution_time	[124.748µs; 124.878µs] or [-0.052%; +0.052%]	None	None	None

Group 4

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`bb7ab76`	1789071525	better-cxx

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_serialize_compressed_pprof_timestamped_x1000	execution_time	953.238µs	957.056µs ± 2.692µs	956.468µs ± 0.818µs	957.379µs	961.650µs	963.923µs	980.184µs	2.48%	3.969	27.178	0.28%	0.190µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_serialize_compressed_pprof_timestamped_x1000	execution_time	[956.683µs; 957.429µs] or [-0.039%; +0.039%]	None	None	None

Group 5

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`bb7ab76`	1789071525	better-cxx

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_add_sample_frames_x1000	execution_time	4.098ms	4.104ms ± 0.003ms	4.104ms ± 0.001ms	4.105ms	4.108ms	4.109ms	4.131ms	0.67%	4.218	34.686	0.07%	0.000ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_add_sample_frames_x1000	execution_time	[4.104ms; 4.104ms] or [-0.010%; +0.010%]	None	None	None

Group 6

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`bb7ab76`	1789071525	better-cxx

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_add_sample2_frames_x1000	execution_time	688.298µs	689.652µs ± 0.491µs	689.654µs ± 0.314µs	689.974µs	690.466µs	690.751µs	691.267µs	0.23%	0.012	0.136	0.07%	0.035µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_add_sample2_frames_x1000	execution_time	[689.584µs; 689.720µs] or [-0.010%; +0.010%]	None	None	None

Baseline

Baseline benchmark details

Group 1

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`8b0fa65`	1789070470	main

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profiles_dictionary/profile_string_inserts/threads/1	execution_time	377.829µs	379.411µs ± 1.215µs	379.461µs ± 0.453µs	379.699µs	381.211µs	382.051µs	382.261µs	0.74%	1.123	1.224	0.30%	0.384µs	1	10
profiles_dictionary/profile_string_inserts/threads/1	throughput	2678796.696op/s	2698947.019op/s ± 8614.571op/s	2698565.431op/s ± 3221.426op/s	2704517.744op/s	2708687.729op/s	2709913.327op/s	2710219.727op/s	0.43%	-1.105	1.190	0.30%	2724.167op/s	1	10
profiles_dictionary/profile_string_inserts/threads/16	execution_time	1.992ms	1.998ms ± 0.006ms	1.995ms ± 0.002ms	2.002ms	2.007ms	2.009ms	2.010ms	0.76%	0.875	-0.523	0.29%	0.002ms	1	10
profiles_dictionary/profile_string_inserts/threads/16	throughput	8151617.328op/s	8200937.774op/s ± 24612.006op/s	8213270.871op/s ± 10299.676op/s	8218355.862op/s	8223596.670op/s	8223784.757op/s	8223831.778op/s	0.13%	-0.869	-0.537	0.28%	7783.000op/s	1	10
profiles_dictionary/profile_string_inserts/threads/2	execution_time	633.227µs	636.033µs ± 3.356µs	634.423µs ± 1.049µs	636.984µs	641.973µs	642.005µs	642.014µs	1.20%	1.098	-0.424	0.50%	1.061µs	1	10
profiles_dictionary/profile_string_inserts/threads/2	throughput	3189963.439op/s	3220040.175op/s ± 16899.659op/s	3228130.623op/s ± 5337.850op/s	3231005.222op/s	3233643.651op/s	3234111.949op/s	3234229.023op/s	0.19%	-1.092	-0.434	0.50%	5344.141op/s	1	10
profiles_dictionary/profile_string_inserts/threads/4	execution_time	815.122µs	824.131µs ± 5.371µs	824.284µs ± 2.503µs	825.760µs	831.924µs	835.074µs	835.862µs	1.40%	0.628	0.845	0.62%	1.698µs	1	10
profiles_dictionary/profile_string_inserts/threads/4	throughput	4900329.649op/s	4970275.275op/s ± 32268.459op/s	4969160.750op/s ± 15080.510op/s	4987354.310op/s	5010214.911op/s	5022054.976op/s	5025014.992op/s	1.12%	-0.582	0.805	0.62%	10204.183op/s	1	10

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profiles_dictionary/profile_string_inserts/threads/1	execution_time	[378.658µs; 380.164µs] or [-0.198%; +0.198%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/1	throughput	[2693607.750op/s; 2704286.287op/s] or [-0.198%; +0.198%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/16	execution_time	[1.994ms; 2.002ms] or [-0.186%; +0.186%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/16	throughput	[8185683.375op/s; 8216192.174op/s] or [-0.186%; +0.186%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/2	execution_time	[633.953µs; 638.113µs] or [-0.327%; +0.327%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/2	throughput	[3209565.851op/s; 3230514.500op/s] or [-0.325%; +0.325%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/4	execution_time	[820.802µs; 827.459µs] or [-0.404%; +0.404%]	None	None	None
profiles_dictionary/profile_string_inserts/threads/4	throughput	[4950275.444op/s; 4990275.105op/s] or [-0.402%; +0.402%]	None	None	None

Group 2

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`8b0fa65`	1789070470	main

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_add_sample_timestamped_x1000	execution_time	4.003ms	4.010ms ± 0.009ms	4.009ms ± 0.002ms	4.011ms	4.014ms	4.016ms	4.134ms	3.12%	12.100	158.771	0.23%	0.001ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_add_sample_timestamped_x1000	execution_time	[4.009ms; 4.011ms] or [-0.032%; +0.032%]	None	None	None

Group 3

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`8b0fa65`	1789070470	main

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
benching string interning on wordpress profile	execution_time	124.693µs	125.764µs ± 0.351µs	125.740µs ± 0.171µs	125.914µs	126.239µs	126.725µs	127.683µs	1.55%	1.432	8.315	0.28%	0.025µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
benching string interning on wordpress profile	execution_time	[125.716µs; 125.813µs] or [-0.039%; +0.039%]	None	None	None

Group 4

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`8b0fa65`	1789070470	main

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_serialize_compressed_pprof_timestamped_x1000	execution_time	934.759µs	937.933µs ± 2.100µs	937.419µs ± 0.889µs	938.432µs	941.713µs	944.371µs	951.152µs	1.46%	2.046	7.532	0.22%	0.148µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_serialize_compressed_pprof_timestamped_x1000	execution_time	[937.642µs; 938.224µs] or [-0.031%; +0.031%]	None	None	None

Group 5

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`8b0fa65`	1789070470	main

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_add_sample_frames_x1000	execution_time	4.006ms	4.014ms ± 0.002ms	4.014ms ± 0.002ms	4.016ms	4.018ms	4.020ms	4.021ms	0.17%	-0.205	0.381	0.06%	0.000ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_add_sample_frames_x1000	execution_time	[4.014ms; 4.014ms] or [-0.009%; +0.009%]	None	None	None

Group 6

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	`8b0fa65`	1789070470	main

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
profile_add_sample2_frames_x1000	execution_time	712.829µs	714.374µs ± 0.620µs	714.354µs ± 0.416µs	714.777µs	715.408µs	715.854µs	716.604µs	0.31%	0.280	0.376	0.09%	0.044µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
profile_add_sample2_frames_x1000	execution_time	[714.288µs; 714.460µs] or [-0.012%; +0.012%]	None	None	None

dd-octo-sts · 2026-09-09T19:43:29Z

Artifact Size Benchmark Report

aarch64-alpine-linux-musl

Artifact	Baseline	Commit	Change
/aarch64-alpine-linux-musl/lib/libdatadog_profiling.a	91.32 MB	91.32 MB	+0% (+784 B) 👌
/aarch64-alpine-linux-musl/lib/libdatadog_profiling.so	8.45 MB	8.45 MB	0% (0 B) 👌

aarch64-unknown-linux-gnu

Artifact	Baseline	Commit	Change
/aarch64-unknown-linux-gnu/lib/libdatadog_profiling.a	102.61 MB	102.61 MB	+0% (+840 B) 👌
/aarch64-unknown-linux-gnu/lib/libdatadog_profiling.so	11.38 MB	11.38 MB	0% (0 B) 👌

libdatadog-x64-windows

Artifact	Baseline	Commit	Change
/libdatadog-x64-windows/debug/dynamic/datadog_profiling_ffi.dll	27.32 MB	27.32 MB	+0% (+1.00 KB) 👌
/libdatadog-x64-windows/debug/dynamic/datadog_profiling_ffi.lib	96.08 KB	96.08 KB	0% (0 B) 👌
/libdatadog-x64-windows/debug/dynamic/datadog_profiling_ffi.pdb	185.38 MB	185.40 MB	+0% (+16.00 KB) 👌
/libdatadog-x64-windows/debug/static/datadog_profiling_ffi.lib	804.37 MB	804.45 MB	+0% (+79.98 KB) 👌
/libdatadog-x64-windows/release/dynamic/datadog_profiling_ffi.dll	9.00 MB	9.00 MB	0% (0 B) 👌
/libdatadog-x64-windows/release/dynamic/datadog_profiling_ffi.lib	96.08 KB	96.08 KB	0% (0 B) 👌
/libdatadog-x64-windows/release/dynamic/datadog_profiling_ffi.pdb	26.20 MB	26.20 MB	0% (0 B) 👌
/libdatadog-x64-windows/release/static/datadog_profiling_ffi.lib	52.21 MB	52.21 MB	-0% (-54 B) 👌

libdatadog-x86-windows

Artifact	Baseline	Commit	Change
/libdatadog-x86-windows/debug/dynamic/datadog_profiling_ffi.dll	23.82 MB	23.82 MB	+0% (+512 B) 👌
/libdatadog-x86-windows/debug/dynamic/datadog_profiling_ffi.lib	97.58 KB	97.58 KB	0% (0 B) 👌
/libdatadog-x86-windows/debug/dynamic/datadog_profiling_ffi.pdb	190.43 MB	190.48 MB	+.02% (+56.00 KB) 🔍
/libdatadog-x86-windows/debug/static/datadog_profiling_ffi.lib	785.15 MB	785.23 MB	+.01% (+82.03 KB) 🔍
/libdatadog-x86-windows/release/dynamic/datadog_profiling_ffi.dll	6.96 MB	6.96 MB	0% (0 B) 👌
/libdatadog-x86-windows/release/dynamic/datadog_profiling_ffi.lib	97.58 KB	97.58 KB	0% (0 B) 👌
/libdatadog-x86-windows/release/dynamic/datadog_profiling_ffi.pdb	28.18 MB	28.18 MB	0% (0 B) 👌
/libdatadog-x86-windows/release/static/datadog_profiling_ffi.lib	49.70 MB	49.70 MB	-0% (-28 B) 👌

x86_64-alpine-linux-musl

Artifact	Baseline	Commit	Change
/x86_64-alpine-linux-musl/lib/libdatadog_profiling.a	81.48 MB	81.48 MB	+0% (+48 B) 👌
/x86_64-alpine-linux-musl/lib/libdatadog_profiling.so	9.41 MB	9.41 MB	0% (0 B) 👌

x86_64-unknown-linux-gnu

Artifact	Baseline	Commit	Change
/x86_64-unknown-linux-gnu/lib/libdatadog_profiling.a	97.21 MB	97.21 MB	+0% (+496 B) 👌
/x86_64-unknown-linux-gnu/lib/libdatadog_profiling.so	11.46 MB	11.46 MB	0% (0 B) 👌

ekump

LGTM for the AGENTS.md change

taegyunkim · 2026-09-11T17:06:17Z

I'm not sure why this file needs to be updated as a part of this PR.

taegyunkim · 2026-09-11T17:33:02Z

+        /// borrowed only for the duration of this call. endtime_ns is an
+        /// optional end timestamp in nanoseconds; pass 0 to record the sample
+        /// without a timestamp.
+        fn add_sample2(self: &mut Profile, sample: &Sample2, endtime_ns: i64) -> Status;


Since libdd_profiling::cxx::ffi is not intended for Rust callers, please make the bridge module and its Rust implementation surface crate-private. The current public visibility accidentally exposes an unsound safe API to downstream Rust, as demonstrated by this reproducer (https://gist.github.com/taegyunkim/8d0b2ff9fe8e95262299fd7ad40c3a1f), which uses #![forbid(unsafe_code)] but still segfaults through add_sample2.

taegyunkim · 2026-09-11T17:41:58Z

Can this also be tested in CI by adding it to .github/workflows/test-ffi.yml?

danielsn added 2 commits September 9, 2026 15:20

feat(profiling): add timestamped samples to CXX profile API

6c9dde7

feat(profiling): expose dictionary-backed samples in CXX API

7aeac76

danielsn commented Sep 9, 2026

View reviewed changes

danielsn marked this pull request as ready for review September 9, 2026 19:18

danielsn requested review from a team as code owners September 9, 2026 19:18

morrisonlevi reviewed Sep 9, 2026

View reviewed changes

chatgpt-codex-connector Bot reviewed Sep 9, 2026

View reviewed changes

feat(profiling): expose split CXX profile export API

ec38be2

danielsn changed the title ~~feat(profiling): expose dictionary-backed samples in CXX API~~ feat(profiling)!: expose missing API in CXX Sep 9, 2026

feat(profiling): normalize CXX error-returning APIs

bb7ab76

danielsn requested a review from a team as a code owner September 10, 2026 20:18

ekump approved these changes Sep 10, 2026

View reviewed changes

taegyunkim reviewed Sep 11, 2026

View reviewed changes

Uh oh!

Conversation

danielsn commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does this PR do?

Motivation

Additional Notes

How to test the change?

Uh oh!

github-actions Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

📚 Documentation Check Results

📦 libdd-profiling-ffi - 770 warning(s)

📦 libdd-profiling - 649 warning(s)

Uh oh!

danielsn Sep 9, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔒 Cargo Deny Results

📦 libdd-profiling-ffi - 4 error(s)

📦 libdd-profiling - 4 error(s)

Uh oh!

datadog-datadog-prod-us1 Bot commented Sep 9, 2026 • edited by datadog-datadog-prod-us1-2 Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

morrisonlevi Sep 9, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Sep 9, 2026

Choose a reason for hiding this comment

Uh oh!

pr-commenter Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Benchmarks

Comparison

Explanation

More details about the CI and significant changes

Candidate

Group 1

Group 2

Group 3

Group 4

Group 5

Group 6

Baseline

Group 1

Group 2

Group 3

Group 4

Group 5

Group 6

Uh oh!

dd-octo-sts Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Artifact Size Benchmark Report

Uh oh!

ekump left a comment

Choose a reason for hiding this comment

Uh oh!

taegyunkim Sep 11, 2026

Choose a reason for hiding this comment

Uh oh!

taegyunkim Sep 11, 2026

Choose a reason for hiding this comment

Uh oh!

taegyunkim Sep 11, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

danielsn commented Sep 9, 2026 •

edited

Loading

github-actions Bot commented Sep 9, 2026 •

edited

Loading

📦 `libdd-profiling-ffi` - 770 warning(s)

📦 `libdd-profiling` - 649 warning(s)

github-actions Bot commented Sep 9, 2026 •

edited

Loading

📦 `libdd-profiling-ffi` - 4 error(s)

📦 `libdd-profiling` - 4 error(s)

datadog-datadog-prod-us1 Bot commented Sep 9, 2026 •

edited by datadog-datadog-prod-us1-2 Bot

Loading

pr-commenter Bot commented Sep 9, 2026 •

edited

Loading

dd-octo-sts Bot commented Sep 9, 2026 •

edited

Loading