feat(runtime): add shared runtime by VianneyRuhlmann · Pull Request #1602 · DataDog/libdatadog

VianneyRuhlmann · 2026-02-18T15:21:28Z

What does this PR do?

Implementing the shared runtime

This PR introduces a shared runtime to be used across common components it offers two main features:

Provide access to a tokio runtime to call async code in blocking functions e.g. TraceExporter::send
Spawn background workers on the runtime which can be paused and restarted on forks
The shared runtime runs the workers using the PausableWorker

Update to the Worker trait

This PR also refactors the Worker trait to make it execute a trigger-run loop. The worker can be paused between two calls to run but not within a run execution. This is the same tradeoff between fork-time impact and worker execution than the one in dd-trace-py. Some extra hooks where introduced to allow porting the existing workers to this new lifecycle (e.g. initial_trigger, shutdown).

Expose ffi

The shared runtime FFI exposes an pointer to the SharedRuntime which is wrapped in an Arc. The pointer used by the users is pointing directly at the runtime and the Arc is only materialized when the pointer is dropped.

Porting the existing workers

The update to the worker trait needs to be applied to existing workers (stats exporter, info fetcher, telemetry)

Stats exporter

The stats exporter doesn't need the cancellation token anymore as the flushing of remaining stats is handled in the shutdown hook. The trace exporter can trigger the flush by stopping the worker with the handle

Info fetcher

The info fetcher uses the on_pause hook to release the reference to the IoStack. This used to be done by the TraceExporter stop_worker but is now handled directly in the worker.

Telemetry worker

The telemetry worker required some refactor to fit in the new Worker trait. The main issue is that the trigger needs to save the next action to be used in the run function. This could be done by passing the return value of trigger to run but I'm not sure this will be a common pattern in workers. If this turns out to be we can reconsider it.

Losing telemetry payloads on shutdown

Currently when the telemetry worker is shut down all actions sent before the stop action are processed. With this refactor it's no longer the cases as the shutdown hook bypasses the actions queued. This is deemed acceptable since there's currently no way to know if an action has been posted before or after the stop, this means any action is already at risk of being ignored.

Using the shared runtime in TraceExporter

This PR also integrates the shared runtime in the Trace Exporter. We can now pass a SharedRuntime to the TraceExporter and the workers will be spawned on it. Fork hooks from the TraceExporter are removed as this is now handled by the SharedRuntime. The TraceExporter also uses the shared runtime to block on async code.

Motivation

Be able to share a common tokio runtime across components and manage fork-safety in a single place.

Additional Notes

This PR is a breaking change especially for language which register "fork-safety" hooks as this should now be handled in the SharedRuntime.

How to test the change?

Describe here in detail how the change can be validated.

pr-commenter · 2026-02-18T15:38:33Z

Benchmarks

Comparison

Benchmark execution time: 2026-03-23 10:04:07

Comparing candidate commit 0f91bb3 in PR branch vianney/implement-shared-runtime with baseline commit ed00b92 in branch main.

Found 7 performance improvements and 18 performance regressions! Performance is the same for 34 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 ----------------------------------'

scenario:benching deserializing traces from msgpack to their internal representation

🟩 execution_time [-2.719ms; -2.196ms] or [-5.236%; -4.229%]

scenario:concentrator/add_spans_to_concentrator

🟩 execution_time [-2.331ms; -2.325ms] or [-17.939%; -17.893%]

scenario:credit_card/is_card_number/ 3782-8224-6310-005

🟥 execution_time [+4.159µs; +4.388µs] or [+5.493%; +5.795%]
🟥 throughput [-726222.817op/s; -687203.843op/s] or [-5.498%; -5.203%]

scenario:credit_card/is_card_number/ 378282246310005

🟥 execution_time [+4.495µs; +4.562µs] or [+6.542%; +6.639%]
🟥 throughput [-905940.657op/s; -893369.532op/s] or [-6.225%; -6.139%]

scenario:credit_card/is_card_number/378282246310005

🟥 execution_time [+5.397µs; +5.479µs] or [+8.322%; +8.449%]
🟥 throughput [-1201260.479op/s; -1184205.307op/s] or [-7.791%; -7.680%]

scenario:credit_card/is_card_number/37828224631000521389798

🟥 execution_time [+6.508µs; +6.547µs] or [+14.250%; +14.337%]
🟥 throughput [-2747963.369op/s; -2729139.866op/s] or [-12.550%; -12.464%]

scenario:credit_card/is_card_number/x371413321323331

🟩 execution_time [-612.638ns; -594.594ns] or [-9.519%; -9.239%]
🟩 throughput [+15836261.717op/s; +16366589.903op/s] or [+10.192%; +10.534%]

scenario:credit_card/is_card_number_no_luhn/ 378282246310005

🟥 execution_time [+3.276µs; +3.327µs] or [+6.038%; +6.132%]
🟥 throughput [-1065513.012op/s; -1048901.258op/s] or [-5.781%; -5.691%]

scenario:credit_card/is_card_number_no_luhn/378282246310005

🟥 execution_time [+4.375µs; +4.435µs] or [+8.682%; +8.799%]
🟥 throughput [-1605155.624op/s; -1584513.420op/s] or [-8.089%; -7.985%]

scenario:credit_card/is_card_number_no_luhn/37828224631000521389798

🟥 execution_time [+6.512µs; +6.532µs] or [+14.259%; +14.304%]
🟥 throughput [-2741213.791op/s; -2731710.897op/s] or [-12.519%; -12.475%]

scenario:credit_card/is_card_number_no_luhn/x371413321323331

🟩 execution_time [-619.255ns; -601.432ns] or [-9.620%; -9.343%]
🟩 throughput [+16032147.451op/s; +16555643.219op/s] or [+10.320%; +10.657%]

scenario:normalization/normalize_name/normalize_name/Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Lo...

🟥 execution_time [+20.512µs; +20.640µs] or [+11.031%; +11.099%]
🟥 throughput [-537361.613op/s; -534155.334op/s] or [-9.992%; -9.933%]

scenario:normalization/normalize_name/normalize_name/good

🟥 execution_time [+697.722ns; +720.760ns] or [+6.868%; +7.095%]
🟥 throughput [-6525409.947op/s; -6320898.800op/s] or [-6.629%; -6.421%]

scenario:write only interface

🟩 execution_time [-2.560µs; -2.161µs] or [-46.109%; -38.933%]

Candidate

Candidate benchmark details

Group 1

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
sdk_test_data/rules-based	execution_time	144.931µs	147.071µs ± 3.175µs	146.653µs ± 0.447µs	147.136µs	148.510µs	156.309µs	184.376µs	25.72%	9.340	99.707	2.15%	0.225µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
sdk_test_data/rules-based	execution_time	[146.631µs; 147.511µs] or [-0.299%; +0.299%]	None	None	None

Group 2

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
write only interface	execution_time	1.170µs	3.191µs ± 1.437µs	2.985µs ± 0.029µs	3.014µs	3.648µs	13.934µs	14.959µs	401.22%	7.402	55.782	44.91%	0.102µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
write only interface	execution_time	[2.992µs; 3.390µs] or [-6.240%; +6.240%]	None	None	None

Group 3

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
two way interface	execution_time	17.426µs	25.373µs ± 10.278µs	17.831µs ± 0.191µs	34.879µs	44.727µs	52.916µs	73.674µs	313.19%	1.185	1.346	40.41%	0.727µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
two way interface	execution_time	[23.948µs; 26.797µs] or [-5.614%; +5.614%]	None	None	None

Group 4

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
receiver_entry_point/report/2598	execution_time	3.402ms	3.434ms ± 0.024ms	3.427ms ± 0.009ms	3.439ms	3.492ms	3.509ms	3.522ms	2.75%	1.592	1.857	0.71%	0.002ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
receiver_entry_point/report/2598	execution_time	[3.431ms; 3.438ms] or [-0.099%; +0.099%]	None	None	None

Group 5

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
redis/obfuscate_redis_string	execution_time	33.307µs	33.698µs ± 0.689µs	33.396µs ± 0.029µs	33.439µs	35.019µs	35.089µs	38.017µs	13.84%	2.382	7.189	2.04%	0.049µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
redis/obfuscate_redis_string	execution_time	[33.603µs; 33.794µs] or [-0.283%; +0.283%]	None	None	None

Group 6

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
tags/replace_trace_tags	execution_time	2.329µs	2.405µs ± 0.023µs	2.405µs ± 0.007µs	2.412µs	2.450µs	2.457µs	2.460µs	2.31%	-0.772	3.136	0.95%	0.002µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
tags/replace_trace_tags	execution_time	[2.402µs; 2.408µs] or [-0.131%; +0.131%]	None	None	None

Group 7

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
normalization/normalize_name/normalize_name/Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Lo...	execution_time	205.873µs	206.529µs ± 0.388µs	206.463µs ± 0.244µs	206.775µs	207.237µs	207.644µs	207.694µs	0.60%	0.730	0.407	0.19%	0.027µs	1	200
normalization/normalize_name/normalize_name/Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Lo...	throughput	4814772.451op/s	4841957.448op/s ± 9094.993op/s	4843482.852op/s ± 5736.886op/s	4847993.935op/s	4855121.979op/s	4857120.349op/s	4857369.897op/s	0.29%	-0.720	0.386	0.19%	643.113op/s	1	200
normalization/normalize_name/normalize_name/bad-name	execution_time	18.572µs	18.686µs ± 0.085µs	18.662µs ± 0.024µs	18.687µs	18.881µs	18.971µs	19.059µs	2.12%	1.881	3.468	0.45%	0.006µs	1	200
normalization/normalize_name/normalize_name/bad-name	throughput	52469769.853op/s	53516206.309op/s ± 241850.870op/s	53584727.456op/s ± 68451.582op/s	53647702.836op/s	53760308.032op/s	53811444.071op/s	53845651.895op/s	0.49%	-1.855	3.350	0.45%	17101.439op/s	1	200
normalization/normalize_name/normalize_name/good	execution_time	10.755µs	10.868µs ± 0.066µs	10.866µs ± 0.040µs	10.902µs	10.963µs	11.032µs	11.334µs	4.31%	1.892	11.078	0.61%	0.005µs	1	200
normalization/normalize_name/normalize_name/good	throughput	88231530.312op/s	92017223.669op/s ± 552297.866op/s	92032869.384op/s ± 337163.893op/s	92396638.625op/s	92772413.524op/s	92948625.692op/s	92979151.377op/s	1.03%	-1.727	9.688	0.60%	39053.357op/s	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
normalization/normalize_name/normalize_name/Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Lo...	execution_time	[206.475µs; 206.583µs] or [-0.026%; +0.026%]	None	None	None
normalization/normalize_name/normalize_name/Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Lo...	throughput	[4840696.969op/s; 4843217.926op/s] or [-0.026%; +0.026%]	None	None	None
normalization/normalize_name/normalize_name/bad-name	execution_time	[18.675µs; 18.698µs] or [-0.063%; +0.063%]	None	None	None
normalization/normalize_name/normalize_name/bad-name	throughput	[53482688.105op/s; 53549724.514op/s] or [-0.063%; +0.063%]	None	None	None
normalization/normalize_name/normalize_name/good	execution_time	[10.859µs; 10.877µs] or [-0.084%; +0.084%]	None	None	None
normalization/normalize_name/normalize_name/good	throughput	[91940680.496op/s; 92093766.841op/s] or [-0.083%; +0.083%]	None	None	None

Group 8

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
credit_card/is_card_number/	execution_time	3.893µs	3.912µs ± 0.003µs	3.912µs ± 0.002µs	3.914µs	3.916µs	3.917µs	3.918µs	0.15%	-1.736	13.437	0.07%	0.000µs	1	200
credit_card/is_card_number/	throughput	255256364.439op/s	255633172.885op/s ± 169487.468op/s	255647860.622op/s ± 104743.766op/s	255731736.312op/s	255835156.437op/s	255887197.760op/s	256891798.119op/s	0.49%	1.761	13.653	0.07%	11984.574op/s	1	200
credit_card/is_card_number/ 3782-8224-6310-005	execution_time	79.440µs	79.988µs ± 0.333µs	79.923µs ± 0.224µs	80.183µs	80.616µs	80.818µs	81.367µs	1.81%	0.848	0.691	0.42%	0.024µs	1	200
credit_card/is_card_number/ 3782-8224-6310-005	throughput	12290020.091op/s	12502164.645op/s ± 51839.113op/s	12512016.541op/s ± 35177.219op/s	12542489.919op/s	12570537.941op/s	12579027.585op/s	12588182.738op/s	0.61%	-0.824	0.608	0.41%	3665.579op/s	1	200
credit_card/is_card_number/ 378282246310005	execution_time	72.729µs	73.244µs ± 0.226µs	73.228µs ± 0.129µs	73.350µs	73.622µs	73.990µs	74.278µs	1.43%	1.119	3.156	0.31%	0.016µs	1	200
credit_card/is_card_number/ 378282246310005	throughput	13462858.479op/s	13653065.936op/s ± 41929.049op/s	13655905.779op/s ± 24024.882op/s	13682151.650op/s	13706611.430op/s	13733842.524op/s	13749654.126op/s	0.69%	-1.083	3.011	0.31%	2964.831op/s	1	200
credit_card/is_card_number/37828224631	execution_time	3.890µs	3.912µs ± 0.003µs	3.911µs ± 0.001µs	3.913µs	3.917µs	3.920µs	3.925µs	0.35%	-0.684	12.561	0.08%	0.000µs	1	200
credit_card/is_card_number/37828224631	throughput	254773465.339op/s	255630159.543op/s ± 209173.962op/s	255662481.545op/s ± 89916.058op/s	255735444.962op/s	255851653.511op/s	255944518.159op/s	257093963.470op/s	0.56%	0.718	12.762	0.08%	14790.833op/s	1	200
credit_card/is_card_number/378282246310005	execution_time	69.700µs	70.292µs ± 0.280µs	70.259µs ± 0.208µs	70.484µs	70.775µs	70.983µs	71.204µs	1.34%	0.518	0.104	0.40%	0.020µs	1	200
credit_card/is_card_number/378282246310005	throughput	14044163.098op/s	14226565.138op/s ± 56488.674op/s	14233030.040op/s ± 42111.947op/s	14269526.389op/s	14305621.837op/s	14326856.661op/s	14347176.385op/s	0.80%	-0.496	0.066	0.40%	3994.352op/s	1	200
credit_card/is_card_number/37828224631000521389798	execution_time	52.132µs	52.197µs ± 0.031µs	52.196µs ± 0.020µs	52.216µs	52.252µs	52.283µs	52.304µs	0.21%	0.557	0.677	0.06%	0.002µs	1	200
credit_card/is_card_number/37828224631000521389798	throughput	19118969.981op/s	19158269.265op/s ± 11225.132op/s	19158678.565op/s ± 7496.842op/s	19166044.956op/s	19174504.333op/s	19179887.931op/s	19182175.922op/s	0.12%	-0.552	0.668	0.06%	793.737op/s	1	200
credit_card/is_card_number/x371413321323331	execution_time	5.693µs	5.832µs ± 0.065µs	5.832µs ± 0.044µs	5.877µs	5.934µs	5.980µs	6.002µs	2.91%	-0.013	-0.444	1.11%	0.005µs	1	200
credit_card/is_card_number/x371413321323331	throughput	166623921.322op/s	171476941.287op/s ± 1910164.473op/s	171471844.403op/s ± 1301374.724op/s	172718381.375op/s	174620591.436op/s	175581248.953op/s	175660400.314op/s	2.44%	0.065	-0.454	1.11%	135069.025op/s	1	200
credit_card/is_card_number_no_luhn/	execution_time	3.898µs	3.912µs ± 0.002µs	3.912µs ± 0.001µs	3.913µs	3.915µs	3.918µs	3.921µs	0.25%	-0.137	7.056	0.06%	0.000µs	1	200
credit_card/is_card_number_no_luhn/	throughput	255010171.984op/s	255632535.071op/s ± 149264.270op/s	255644972.421op/s ± 85881.507op/s	255725853.886op/s	255809370.017op/s	255897874.062op/s	256514201.728op/s	0.34%	0.153	7.109	0.06%	10554.578op/s	1	200
credit_card/is_card_number_no_luhn/ 3782-8224-6310-005	execution_time	64.147µs	64.378µs ± 0.165µs	64.327µs ± 0.072µs	64.432µs	64.716µs	64.871µs	65.500µs	1.82%	2.547	10.812	0.26%	0.012µs	1	200
credit_card/is_card_number_no_luhn/ 3782-8224-6310-005	throughput	15267284.646op/s	15533252.939op/s ± 39529.668op/s	15545548.607op/s ± 17360.312op/s	15558820.082op/s	15570589.959op/s	15576595.327op/s	15589299.505op/s	0.28%	-2.499	10.382	0.25%	2795.170op/s	1	200
credit_card/is_card_number_no_luhn/ 378282246310005	execution_time	57.347µs	57.561µs ± 0.118µs	57.539µs ± 0.067µs	57.615µs	57.770µs	57.950µs	58.025µs	0.84%	1.306	2.238	0.20%	0.008µs	1	200
credit_card/is_card_number_no_luhn/ 378282246310005	throughput	17234065.250op/s	17372903.239op/s ± 35565.010op/s	17379462.691op/s ± 20291.973op/s	17396800.729op/s	17415629.923op/s	17425135.481op/s	17437843.608op/s	0.34%	-1.290	2.181	0.20%	2514.826op/s	1	200
credit_card/is_card_number_no_luhn/37828224631	execution_time	3.894µs	3.912µs ± 0.003µs	3.912µs ± 0.002µs	3.913µs	3.915µs	3.918µs	3.920µs	0.23%	-1.168	12.070	0.06%	0.000µs	1	200
credit_card/is_card_number_no_luhn/37828224631	throughput	255076974.490op/s	255644208.788op/s ± 165894.311op/s	255654378.488op/s ± 100274.325op/s	255746360.425op/s	255849162.620op/s	255904554.747op/s	256819579.274op/s	0.46%	1.193	12.237	0.06%	11730.499op/s	1	200
credit_card/is_card_number_no_luhn/378282246310005	execution_time	54.402µs	54.801µs ± 0.185µs	54.768µs ± 0.117µs	54.909µs	55.162µs	55.345µs	55.451µs	1.25%	0.875	0.864	0.34%	0.013µs	1	200
credit_card/is_card_number_no_luhn/378282246310005	throughput	18033989.636op/s	18247909.075op/s ± 61272.938op/s	18258717.041op/s ± 39081.013op/s	18291157.609op/s	18329508.677op/s	18353322.354op/s	18381542.360op/s	0.67%	-0.854	0.815	0.33%	4332.651op/s	1	200
credit_card/is_card_number_no_luhn/37828224631000521389798	execution_time	52.116µs	52.190µs ± 0.032µs	52.187µs ± 0.018µs	52.206µs	52.249µs	52.287µs	52.306µs	0.23%	0.820	1.275	0.06%	0.002µs	1	200
credit_card/is_card_number_no_luhn/37828224631000521389798	throughput	19118139.948op/s	19160724.965op/s ± 11642.802op/s	19161977.845op/s ± 6497.681op/s	19168213.830op/s	19177005.141op/s	19181852.521op/s	19187982.567op/s	0.14%	-0.815	1.265	0.06%	823.270op/s	1	200
credit_card/is_card_number_no_luhn/x371413321323331	execution_time	5.696µs	5.827µs ± 0.064µs	5.825µs ± 0.043µs	5.872µs	5.934µs	5.979µs	6.016µs	3.28%	0.244	-0.215	1.10%	0.005µs	1	200
credit_card/is_card_number_no_luhn/x371413321323331	throughput	166221690.520op/s	171646408.860op/s ± 1886445.451op/s	171678942.084op/s ± 1268577.167op/s	172877957.639op/s	174608001.924op/s	175503585.983op/s	175565404.842op/s	2.26%	-0.187	-0.261	1.10%	133391.837op/s	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
credit_card/is_card_number/	execution_time	[3.911µs; 3.912µs] or [-0.009%; +0.009%]	None	None	None
credit_card/is_card_number/	throughput	[255609683.552op/s; 255656662.218op/s] or [-0.009%; +0.009%]	None	None	None
credit_card/is_card_number/ 3782-8224-6310-005	execution_time	[79.941µs; 80.034µs] or [-0.058%; +0.058%]	None	None	None
credit_card/is_card_number/ 3782-8224-6310-005	throughput	[12494980.243op/s; 12509349.048op/s] or [-0.057%; +0.057%]	None	None	None
credit_card/is_card_number/ 378282246310005	execution_time	[73.213µs; 73.276µs] or [-0.043%; +0.043%]	None	None	None
credit_card/is_card_number/ 378282246310005	throughput	[13647254.973op/s; 13658876.899op/s] or [-0.043%; +0.043%]	None	None	None
credit_card/is_card_number/37828224631	execution_time	[3.911µs; 3.912µs] or [-0.011%; +0.011%]	None	None	None
credit_card/is_card_number/37828224631	throughput	[255601170.044op/s; 255659149.042op/s] or [-0.011%; +0.011%]	None	None	None
credit_card/is_card_number/378282246310005	execution_time	[70.253µs; 70.331µs] or [-0.055%; +0.055%]	None	None	None
credit_card/is_card_number/378282246310005	throughput	[14218736.351op/s; 14234393.924op/s] or [-0.055%; +0.055%]	None	None	None
credit_card/is_card_number/37828224631000521389798	execution_time	[52.193µs; 52.201µs] or [-0.008%; +0.008%]	None	None	None
credit_card/is_card_number/37828224631000521389798	throughput	[19156713.570op/s; 19159824.960op/s] or [-0.008%; +0.008%]	None	None	None
credit_card/is_card_number/x371413321323331	execution_time	[5.823µs; 5.841µs] or [-0.154%; +0.154%]	None	None	None
credit_card/is_card_number/x371413321323331	throughput	[171212210.862op/s; 171741671.711op/s] or [-0.154%; +0.154%]	None	None	None
credit_card/is_card_number_no_luhn/	execution_time	[3.912µs; 3.912µs] or [-0.008%; +0.008%]	None	None	None
credit_card/is_card_number_no_luhn/	throughput	[255611848.479op/s; 255653221.663op/s] or [-0.008%; +0.008%]	None	None	None
credit_card/is_card_number_no_luhn/ 3782-8224-6310-005	execution_time	[64.356µs; 64.401µs] or [-0.035%; +0.035%]	None	None	None
credit_card/is_card_number_no_luhn/ 3782-8224-6310-005	throughput	[15527774.507op/s; 15538731.371op/s] or [-0.035%; +0.035%]	None	None	None
credit_card/is_card_number_no_luhn/ 378282246310005	execution_time	[57.545µs; 57.578µs] or [-0.028%; +0.028%]	None	None	None
credit_card/is_card_number_no_luhn/ 378282246310005	throughput	[17367974.271op/s; 17377832.208op/s] or [-0.028%; +0.028%]	None	None	None
credit_card/is_card_number_no_luhn/37828224631	execution_time	[3.911µs; 3.912µs] or [-0.009%; +0.009%]	None	None	None
credit_card/is_card_number_no_luhn/37828224631	throughput	[255621217.432op/s; 255667200.144op/s] or [-0.009%; +0.009%]	None	None	None
credit_card/is_card_number_no_luhn/378282246310005	execution_time	[54.776µs; 54.827µs] or [-0.047%; +0.047%]	None	None	None
credit_card/is_card_number_no_luhn/378282246310005	throughput	[18239417.235op/s; 18256400.915op/s] or [-0.047%; +0.047%]	None	None	None
credit_card/is_card_number_no_luhn/37828224631000521389798	execution_time	[52.186µs; 52.195µs] or [-0.008%; +0.008%]	None	None	None
credit_card/is_card_number_no_luhn/37828224631000521389798	throughput	[19159111.385op/s; 19162338.545op/s] or [-0.008%; +0.008%]	None	None	None
credit_card/is_card_number_no_luhn/x371413321323331	execution_time	[5.818µs; 5.836µs] or [-0.153%; +0.153%]	None	None	None
credit_card/is_card_number_no_luhn/x371413321323331	throughput	[171384965.664op/s; 171907852.057op/s] or [-0.152%; +0.152%]	None	None	None

Group 9

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
normalization/normalize_trace/test_trace	execution_time	247.686ns	257.827ns ± 11.480ns	253.000ns ± 3.488ns	261.922ns	283.993ns	287.536ns	288.611ns	14.08%	1.315	0.346	4.44%	0.812ns	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
normalization/normalize_trace/test_trace	execution_time	[256.236ns; 259.418ns] or [-0.617%; +0.617%]	None	None	None

Group 10

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
sql/obfuscate_sql_string	execution_time	85.903µs	86.066µs ± 0.158µs	86.040µs ± 0.042µs	86.101µs	86.196µs	86.389µs	87.983µs	2.26%	9.187	107.580	0.18%	0.011µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
sql/obfuscate_sql_string	execution_time	[86.044µs; 86.088µs] or [-0.025%; +0.025%]	None	None	None

Group 11

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

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	161.561µs	162.401µs ± 0.271µs	162.378µs ± 0.110µs	162.494µs	162.860µs	163.399µs	163.718µs	0.83%	1.347	5.454	0.17%	0.019µs	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	[162.363µs; 162.438µs] or [-0.023%; +0.023%]	None	None	None

Group 12

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

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	714.452µs	716.084µs ± 0.711µs	716.015µs ± 0.376µs	716.405µs	717.457µs	718.403µs	718.880µs	0.40%	0.977	2.037	0.10%	0.050µs	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	[715.985µs; 716.182µs] or [-0.014%; +0.014%]	None	None	None

Group 13

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
benching deserializing traces from msgpack to their internal representation	execution_time	49.053ms	49.472ms ± 1.346ms	49.290ms ± 0.092ms	49.393ms	49.551ms	59.488ms	61.443ms	24.66%	7.862	61.244	2.71%	0.095ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
benching deserializing traces from msgpack to their internal representation	execution_time	[49.286ms; 49.659ms] or [-0.377%; +0.377%]	None	None	None

Group 14

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
ip_address/quantize_peer_ip_address_benchmark	execution_time	5.012µs	5.088µs ± 0.049µs	5.101µs ± 0.040µs	5.118µs	5.171µs	5.174µs	5.179µs	1.52%	0.126	-1.226	0.96%	0.003µs	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
ip_address/quantize_peer_ip_address_benchmark	execution_time	[5.081µs; 5.095µs] or [-0.134%; +0.134%]	None	None	None

Group 15

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
normalization/normalize_service/normalize_service/A0000000000000000000000000000000000000000000000000...	execution_time	493.251µs	494.064µs ± 0.410µs	494.036µs ± 0.202µs	494.248µs	494.574µs	495.000µs	497.725µs	0.75%	3.682	30.389	0.08%	0.029µs	1	200
normalization/normalize_service/normalize_service/A0000000000000000000000000000000000000000000000000...	throughput	2009142.940op/s	2024029.056op/s ± 1673.178op/s	2024145.355op/s ± 829.333op/s	2024966.829op/s	2026093.689op/s	2026848.169op/s	2027363.779op/s	0.16%	-3.635	29.856	0.08%	118.312op/s	1	200
normalization/normalize_service/normalize_service/Data🐨dog🐶 繋がっ⛰てて	execution_time	370.123µs	370.646µs ± 0.249µs	370.630µs ± 0.166µs	370.809µs	371.081µs	371.355µs	371.421µs	0.21%	0.395	0.117	0.07%	0.018µs	1	200
normalization/normalize_service/normalize_service/Data🐨dog🐶 繋がっ⛰てて	throughput	2692365.412op/s	2697996.909op/s ± 1814.700op/s	2698107.202op/s ± 1208.352op/s	2699273.816op/s	2700938.317op/s	2701434.854op/s	2701805.487op/s	0.14%	-0.391	0.111	0.07%	128.319op/s	1	200
normalization/normalize_service/normalize_service/Test Conversion 0f Weird !@#$%^&**() Characters	execution_time	167.310µs	167.567µs ± 0.168µs	167.548µs ± 0.113µs	167.661µs	167.854µs	168.074µs	168.210µs	0.39%	0.940	1.058	0.10%	0.012µs	1	200
normalization/normalize_service/normalize_service/Test Conversion 0f Weird !@#$%^&**() Characters	throughput	5944958.806op/s	5967756.940op/s ± 5986.323op/s	5968424.398op/s ± 4024.460op/s	5972447.641op/s	5975634.226op/s	5976715.617op/s	5976945.670op/s	0.14%	-0.933	1.037	0.10%	423.297op/s	1	200
normalization/normalize_service/normalize_service/[empty string]	execution_time	37.160µs	37.248µs ± 0.047µs	37.242µs ± 0.032µs	37.280µs	37.330µs	37.363µs	37.412µs	0.46%	0.590	-0.033	0.13%	0.003µs	1	200
normalization/normalize_service/normalize_service/[empty string]	throughput	26729688.290op/s	26846785.409op/s ± 33797.482op/s	26851659.421op/s ± 22871.999op/s	26872345.403op/s	26892369.663op/s	26902641.244op/s	26910991.188op/s	0.22%	-0.584	-0.046	0.13%	2389.843op/s	1	200
normalization/normalize_service/normalize_service/test_ASCII	execution_time	45.463µs	45.711µs ± 0.219µs	45.642µs ± 0.135µs	45.892µs	46.039µs	46.115µs	47.048µs	3.08%	1.415	5.100	0.48%	0.016µs	1	200
normalization/normalize_service/normalize_service/test_ASCII	throughput	21254951.775op/s	21877102.087op/s ± 104281.329op/s	21909761.212op/s ± 65039.467op/s	21965147.335op/s	21981321.124op/s	21994373.546op/s	21995783.905op/s	0.39%	-1.345	4.542	0.48%	7373.803op/s	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
normalization/normalize_service/normalize_service/A0000000000000000000000000000000000000000000000000...	execution_time	[494.008µs; 494.121µs] or [-0.011%; +0.011%]	None	None	None
normalization/normalize_service/normalize_service/A0000000000000000000000000000000000000000000000000...	throughput	[2023797.170op/s; 2024260.942op/s] or [-0.011%; +0.011%]	None	None	None
normalization/normalize_service/normalize_service/Data🐨dog🐶 繋がっ⛰てて	execution_time	[370.611µs; 370.680µs] or [-0.009%; +0.009%]	None	None	None
normalization/normalize_service/normalize_service/Data🐨dog🐶 繋がっ⛰てて	throughput	[2697745.409op/s; 2698248.409op/s] or [-0.009%; +0.009%]	None	None	None
normalization/normalize_service/normalize_service/Test Conversion 0f Weird !@#$%^&**() Characters	execution_time	[167.544µs; 167.591µs] or [-0.014%; +0.014%]	None	None	None
normalization/normalize_service/normalize_service/Test Conversion 0f Weird !@#$%^&**() Characters	throughput	[5966927.293op/s; 5968586.587op/s] or [-0.014%; +0.014%]	None	None	None
normalization/normalize_service/normalize_service/[empty string]	execution_time	[37.242µs; 37.255µs] or [-0.017%; +0.017%]	None	None	None
normalization/normalize_service/normalize_service/[empty string]	throughput	[26842101.403op/s; 26851469.415op/s] or [-0.017%; +0.017%]	None	None	None
normalization/normalize_service/normalize_service/test_ASCII	execution_time	[45.681µs; 45.741µs] or [-0.066%; +0.066%]	None	None	None
normalization/normalize_service/normalize_service/test_ASCII	throughput	[21862649.698op/s; 21891554.477op/s] or [-0.066%; +0.066%]	None	None	None

Group 16

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
single_flag_killswitch/rules-based	execution_time	190.781ns	192.930ns ± 2.174ns	192.791ns ± 1.353ns	193.695ns	195.698ns	203.059ns	206.532ns	7.13%	2.870	13.239	1.12%	0.154ns	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
single_flag_killswitch/rules-based	execution_time	[192.629ns; 193.231ns] or [-0.156%; +0.156%]	None	None	None

Group 17

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

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	4.141ms	4.145ms ± 0.007ms	4.144ms ± 0.001ms	4.145ms	4.150ms	4.172ms	4.226ms	1.98%	8.026	80.449	0.17%	0.001ms	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	[4.144ms; 4.146ms] or [-0.024%; +0.024%]	None	None	None

Group 18

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
concentrator/add_spans_to_concentrator	execution_time	10.640ms	10.665ms ± 0.015ms	10.663ms ± 0.009ms	10.672ms	10.689ms	10.704ms	10.754ms	0.85%	1.790	7.178	0.14%	0.001ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
concentrator/add_spans_to_concentrator	execution_time	[10.663ms; 10.667ms] or [-0.020%; +0.020%]	None	None	None

Group 19

cpu_model	git_commit_sha	git_commit_date	git_branch
Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz	`0f91bb3`	1774259263	vianney/implement-shared-runtime

scenario	metric	min	mean ± sd	median ± mad	p75	p95	p99	max	peak_to_median_ratio	skewness	kurtosis	cv	sem	runs	sample_size
benching serializing traces from their internal representation to msgpack	execution_time	14.857ms	14.904ms ± 0.031ms	14.897ms ± 0.012ms	14.913ms	14.969ms	15.012ms	15.094ms	1.32%	2.473	9.355	0.20%	0.002ms	1	200

scenario	metric	95% CI mean	Shapiro-Wilk pvalue	Ljung-Box pvalue (lag=1)	Dip test pvalue
benching serializing traces from their internal representation to msgpack	execution_time	[14.899ms; 14.908ms] or [-0.028%; +0.028%]	None	None	None

Baseline

Omitted due to size.

github-actions · 2026-02-20T16:15:28Z

📚 Documentation Check Results

⚠️ 1445 documentation warning(s) found

📦 `libdd-common` - 168 warning(s)

📦 `libdd-data-pipeline` - 799 warning(s)

📦 `libdd-telemetry` - 478 warning(s)

Updated: 2026-03-11 12:57:10 UTC | Commit: 85e0696 | missing-docs job results

github-actions · 2026-02-20T16:16:49Z

Clippy Allow Annotation Report

Comparing clippy allow annotations between branches:

Base Branch: origin/main
PR Branch: origin/vianney/implement-shared-runtime

Summary by Rule

Rule	Base Branch	PR Branch	Change
unwrap_used	18	18	No change (0%)
Total	18	18	No change (0%)

Annotation Counts by File

File	Base Branch	PR Branch	Change
`libdd-data-pipeline/src/stats_exporter.rs`	1	1	No change (0%)
`libdd-data-pipeline/src/telemetry/mod.rs`	1	1	No change (0%)
`libdd-data-pipeline/src/trace_exporter/mod.rs`	2	2	No change (0%)
`libdd-telemetry/src/worker/mod.rs`	13	13	No change (0%)
`libdd-telemetry/src/worker/store.rs`	1	1	No change (0%)

Annotation Stats by Crate

Crate	Base Branch	PR Branch	Change
`clippy-annotation-reporter`	5	5	No change (0%)
`datadog-ffe-ffi`	1	1	No change (0%)
`datadog-ipc`	28	28	No change (0%)
`datadog-live-debugger`	6	6	No change (0%)
`datadog-live-debugger-ffi`	10	10	No change (0%)
`datadog-profiling-replayer`	4	4	No change (0%)
`datadog-remote-config`	3	3	No change (0%)
`datadog-sidecar`	59	59	No change (0%)
`libdd-common`	10	10	No change (0%)
`libdd-common-ffi`	12	12	No change (0%)
`libdd-data-pipeline`	5	5	No change (0%)
`libdd-ddsketch`	2	2	No change (0%)
`libdd-dogstatsd-client`	1	1	No change (0%)
`libdd-profiling`	13	13	No change (0%)
`libdd-telemetry`	19	19	No change (0%)
`libdd-tinybytes`	4	4	No change (0%)
`libdd-trace-normalization`	2	2	No change (0%)
`libdd-trace-obfuscation`	9	9	No change (0%)
`libdd-trace-utils`	15	15	No change (0%)
Total	208	208	No change (0%)

About This Report

This report tracks Clippy allow annotations for specific rules, showing how they've changed in this PR. Decreasing the number of these annotations generally improves code quality.

github-actions · 2026-02-20T16:18:19Z

🔒 Cargo Deny Results

⚠️ 1 issue(s) found, showing only errors (advisories, bans, sources)

📦 `libdd-common` - ✅ No issues

📦 `libdd-data-pipeline` - 1 error(s)

Show output

error[vulnerability]: Denial of Service via Stack Exhaustion
    ┌─ /home/runner/work/libdatadog/libdatadog/Cargo.lock:293:1
    │
293 │ time 0.3.41 registry+https://github.com/rust-lang/crates.io-index
    │ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ security vulnerability detected
    │
    ├ ID: RUSTSEC-2026-0009
    ├ Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0009
    ├ ## Impact
      
      When user-provided input is provided to any type that parses with the RFC 2822 format, a denial of
      service attack via stack exhaustion is possible. The attack relies on formally deprecated and
      rarely-used features that are part of the RFC 2822 format used in a malicious manner. Ordinary,
      non-malicious input will never encounter this scenario.
      
      ## Patches
      
      A limit to the depth of recursion was added in v0.3.47. From this version, an error will be returned
      rather than exhausting the stack.
      
      ## Workarounds
      
      Limiting the length of user input is the simplest way to avoid stack exhaustion, as the amount of
      the stack consumed would be at most a factor of the length of the input.
    ├ Announcement: https://github.com/time-rs/time/blob/main/CHANGELOG.md#0347-2026-02-05
    ├ Solution: Upgrade to >=0.3.47 (try `cargo update -p time`)
    ├ time v0.3.41
      └── tracing-appender v0.2.3
          └── libdd-log v1.0.0
              └── (dev) libdd-data-pipeline v2.0.0

advisories FAILED, bans ok, sources ok

📦 `libdd-telemetry` - ✅ No issues

Updated: 2026-03-11 12:59:51 UTC | Commit: 85e0696 | dependency-check job results

With the refactor the stop action is processed right away without processing all actions sent before it. This can cause some message to be lost if they were submitted right before shutdown.

ekump · 2026-03-11T12:54:17Z

libdd-shared-runtime/src/worker/shared_runtime.rs

+    ///
+    /// # Errors
+    /// Returns an error if it fails to create a runtime.
+    pub fn runtime(&self) -> Result<Arc<Runtime>, io::Error> {


Is returning a single-threaded runtime ok?

When would the underlying runtime not be available? In between the before_fork() and after_fork_* functions?

If a component blocking function (e.g. TraceExporter::send) requires a tokio runtime while we are preparing for a fork, we want to allow the component to access one to avoid losing data (e.g. the trace payload) without restarting the full runtime which we would have to shutdown again. The components should use the runtime with block_on so a current thread runtime will behave in the same way as the multi-threaded one. Since the SharedRuntime doesn't hold a reference to this new runtime, it will be dropped when the blocking function returns guarantying no runtime is running across the fork.

I don't know if it's worth the trouble, but one possibility to avoid misuse here could be to revert the control and take a future and block on it, instead of providing a full runtime to use, if that's the intended use (so a block_on wrapper, mostly). This would constrain what the caller can actually do.

ekump · 2026-03-12T13:36:33Z

libdd-shared-runtime/src/worker/shared_runtime.rs

+}
+
+#[derive(Debug)]
+pub enum WorkerHandleError {


Should you impl std::error::Error for consistency with the other errors?

ekump · 2026-03-12T20:03:31Z

libdd-data-pipeline/src/agent_info/fetcher.rs

+    }
+
+    fn reset(&mut self) {
+        // Drain all messages from the channel to remove messages sent to release the reference on


Suggested change

// Drain all messages from the channel to remove messages sent to release the reference on

// Drain all messages from the channel to remove messages sent to release the reference on IoStack

ekump · 2026-03-12T20:15:00Z

libdd-common/src/worker.rs

-    /// Main worker loop
-    fn run(&mut self) -> impl std::future::Future<Output = ()> + Send;
+/// # Lifecycle
+/// The worker's `Self::run` method should be executed everytime the `Self::trigger` method returns.


Defining a bunch of rules in comments that are not enforced in code is a bit of a smell. Do we anticipate there being more than PausableWorker and perhaps something that is WASM compatible? If there is a small number of impls, probably not a big deal because it looks like the rules are enforced internally in the impl, and not something the consumer has to deal with.

I thought of this comments to be aimed at people implementing the Worker trait, rather than for someone designing a Worker "runner" which for now is supposed to be only PausableWorker. Regarding WASM compatibility where using mostly std async stuff which should work the same, I'm not sure about the JoinHandle which are tokio specific but since node doesn't fork we may also provide an alternative implementation which doesn't support pausing the workers.

To add to the original comment: I believe it's not entirely clear from the wording if this is a constraint on the user/caller, or a guarantee that the implementer/callee needs to uphold.

ekump · 2026-03-12T20:45:54Z

libdd-shared-runtime-ffi/src/shared_runtime.rs

None of these functions are using catch_panic. Is that intentional?

ekump · 2026-03-12T20:49:25Z

libdd-shared-runtime/src/worker/shared_runtime.rs

Why does this live in libdd-data-pipeline? If we don't want a dedicated crate, shouldn't it go in libdd-common?

You're right I'll move it to ddcommon

ekump · 2026-03-12T20:53:16Z

libdd-shared-runtime/src/worker/pausable_worker.rs

Applies to the whole PR: Should we pepper in some debug log messages? Debugging this kind of stuff is tricky.

yannham

I'm still missing a lot of context, so I can't really comment on the integration of the shared runtime and apologies in advance if some remarks or questions aren't relevant. Though the general motivation, the approach and the architecture do look very sensible.

yannham · 2026-03-13T16:12:08Z

libdd-common/src/worker.rs

-    /// Main worker loop
-    fn run(&mut self) -> impl std::future::Future<Output = ()> + Send;
+/// # Lifecycle
+/// The worker's `Self::run` method should be executed everytime the `Self::trigger` method returns.


To add to the original comment: I believe it's not entirely clear from the wording if this is a constraint on the user/caller, or a guarantee that the implementer/callee needs to uphold.

yannham · 2026-03-13T16:22:25Z

libdd-shared-runtime-ffi/src/shared_runtime.rs

+#[repr(C)]
+pub struct SharedRuntimeFFIError {
+    pub code: SharedRuntimeErrorCode,
+    pub msg: *mut c_char,


Minor: can msg be null here to denote an empty one, or is it guaranteed to always be defined? If the latter, maybe mention it.

The msg is generated from the rust error to_string method so it should never be empty. I'll add a comment.

yannham · 2026-03-13T16:48:58Z

libdd-data-pipeline/src/agent_info/fetcher.rs

    info_endpoint: Endpoint,
    refresh_interval: Duration,
    trigger_rx: Option<mpsc::Receiver<()>>,
+    trigger_tx: mpsc::Sender<()>,


naive question, but why does the agent info fetcher has both ends of the trigger channel? Is it something related to the phantom waker issue?

The fetcher needs to send a message to the channel when paused to release the reference to the waker. Otherwise the waker prevents us from dropping the runtime's IoStack which causes issues on forks. @paullegranddc has a better solution here to which we should migrate at some point.

yannham · 2026-03-13T17:05:10Z

libdd-data-pipeline/src/trace_exporter/mod.rs

-                    sleep(Duration::from_millis(100)).await;
-                })
+            exporter.shared_runtime.runtime().unwrap().block_on(async {
+                sleep(Duration::from_millis(100)).await;


Out of curiosity, why does this test use an async runtime to block on a sleep?

yannham · 2026-03-13T17:10:32Z

libdd-data-pipeline/src/trace_exporter/stats.rs

+    let worker_handle = ctx
+        .shared_runtime
+        .spawn_worker(stats_exporter)
+        .map_err(|e| anyhow::anyhow!(e.to_string()))?;


Very minor: if e impl Display, anyhow! already handles the conversion I believe.

Suggested change

.map_err(|e| anyhow::anyhow!(e.to_string()))?;

.map_err(|e| anyhow::anyhow!(e))?;

yannham · 2026-03-16T10:42:30Z

libdd-data-pipeline/src/shared_runtime.rs

+    /// This allows external code to spawn additional tasks on the runtime if needed.
+    ///
+    /// # Warning
+    /// Since this method can return a single-threaded runtime it should only be use to


Suggested change

/// Since this method can return a single-threaded runtime it should only be use to

/// Since this method can return a single-threaded runtime it should only be used to

yannham · 2026-03-16T10:44:33Z

libdd-shared-runtime/src/worker/shared_runtime.rs

+    ///
+    /// # Errors
+    /// Returns an error if it fails to create a runtime.
+    pub fn runtime(&self) -> Result<Arc<Runtime>, io::Error> {


I don't know if it's worth the trouble, but one possibility to avoid misuse here could be to revert the control and take a future and block on it, instead of providing a full runtime to use, if that's the intended use (so a block_on wrapper, mostly). This would constrain what the caller can actually do.

yannham · 2026-03-16T10:48:24Z

libdd-shared-runtime/src/worker/shared_runtime.rs

+    /// This function should not take ownership of the SharedRuntime as it will cause the runtime
+    /// to be dropped in a non-blocking context causing a panic.


I fear I don't get this comment. Is it something the caller must avoid? Or is it a statement, like "this function doesn't take ownership"? But then I'm not sure what does "this function takes ownership" means in this context, since it's an concrete implementation taking a &self.

yannham · 2026-03-16T11:06:52Z

libdd-data-pipeline/src/pausable_worker.rs

 /// # Time-to-pause
 /// This loop should yield regularly to reduce time-to-pause. See [`tokio::task::yield_now`].
 ///
 /// # Cancellation safety


Just to be sure: this requires the inner worker's various async function to return cancel-safe futures, right? Is this something any user of PausableWorker has to ensure when wrapping a worker into a pausable worker? If yes, maybe it could make sense to repeat it in the documentation of new().

As an aside, this sounds like a complicated constraint to enforce, as a user of PausableWorker: you basically have to carefully look at the implementation of the worker you use and make sure it's indeed cancel-safe, and any future change to this implementation could invalidate that assumption. It's just a random idea, but another approach could be to push the burden of the proof to the Worker implementers instead (Make a CanBePausedWorkerButWithABetterName: Worker marker trait, and have T: CanBePaused... + Send + Sync + 'static instead?), which allegedly know better if they are cancel safe. Or just add this constraint in Worker itself, requiring all implementers to uphold it, if in practice we expect all workers to be pausable.

yannham · 2026-03-16T11:17:07Z

libdd-telemetry/src/worker/mod.rs

+        let (worker_handle, worker) = self.build_worker(tokio_runtime.clone());

-        let join_handle = tokio_runtime.spawn(async move { worker.run().await });
+        let join_handle = tokio_runtime.spawn(async move { worker.run_loop().await });


I'm probably missing context here, but why does this use the current tokio runtime directly instead of the shared runtime?

VianneyRuhlmann added 11 commits February 20, 2026 17:03

feat(worker): Add trigger to worker trait

0e6fc77

feat(data_pipeline): add SharedRuntime

305b853

feat(worker): add initial trigger

66e9e06

feat(agent_info): use initial trigger

b672262

feat(stats): implement stats worker

23022a6

fix(shared_runtime): fix compile error

6c9c7f3

data-pipeline: pause all workers before joining in before_fork

acc69ae

feat(exporter): use shared runtime in trace exporter

17a9bb8

feat(shared_runtime): add worker handle

238a430

refactor(worker): remove stopped status

592c24c

chore(telemetry): move telemetry shutdown to worker

ff6448c

VianneyRuhlmann force-pushed the vianney/implement-shared-runtime branch from f615fad to ff6448c Compare February 20, 2026 16:13

github-actions bot added telemetry common data-pipeline labels Feb 20, 2026

VianneyRuhlmann added 11 commits February 23, 2026 14:24

chore(shared-runtime): update error types

0dbce86

chore(shared-runtime): return detailed errors in before_fork

58bff21

chore(runtime): doc

16ab63b

test(telemetry): use shared runtime in tests

3625c0a

test(telemetry): use client builder

f86890a

chore(runtime): fix nit

3d285ae

refactor(runtime): move shutdown to runtime

1f8c0cf

test(telemetry): Add sleep after send

62a63d6

With the refactor the stop action is processed right away without processing all actions sent before it. This can cause some message to be lost if they were submitted right before shutdown.

test(telemetry): fix deadlocks in telemetry tests

c1270f1

refactor(runtime): skip shutdown when runtime is None

608556e

feat(runtime): add runtime to builder

34ec7d9

VianneyRuhlmann added the breaking-change label Mar 9, 2026

VianneyRuhlmann added 4 commits March 9, 2026 14:57

docs(runtime): add warnings

3119d5f

feat(shared_runtime): add shared runtime ffi

74425ba

feat(trace_exporter): add shutdown method

2869d39

fix(shared_runtime): add on_pause hook to release waker in info fetcher

4b1d0b6

This comment has been minimized.

Sign in to view

VianneyRuhlmann added 3 commits March 10, 2026 16:59

fix(telemetry): add reset hook to telemetry

0e74cad

fix(telemetry): fix spawn and run loop for telemetry

544b840

docs(info_fetcher): update doc for running the fetcher

a3e2c38

VianneyRuhlmann marked this pull request as ready for review March 10, 2026 17:10

VianneyRuhlmann requested review from a team as code owners March 10, 2026 17:10

VianneyRuhlmann added 3 commits March 11, 2026 11:31

feat(runtime-ffi): remove redundant allocation of Box

67fcf2a

Merge branch 'main' into vianney/implement-shared-runtime

018028f

format

6420f28

ekump reviewed Mar 11, 2026

View reviewed changes

feat(runtime-ffi): use new handle in trace exporter builder

72f61f6

ekump reviewed Mar 12, 2026

View reviewed changes

libdd-shared-runtime-ffi/src/shared_runtime.rs

Copy link

Contributor

ekump Mar 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

None of these functions are using catch_panic. Is that intentional?

ekump reviewed Mar 12, 2026

View reviewed changes

VianneyRuhlmann added 2 commits March 13, 2026 13:58

chore: catch panics on ffi

7bac719

feat(telemetry): add check for existing action

fc1c81b

yannham reviewed Mar 16, 2026

View reviewed changes

VianneyRuhlmann added 3 commits March 20, 2026 15:52

refactor: move shared runtime to a separate crate

57de7d4

chore: apply suggestions

fb187ce

chore(codeowners): add libdd-shared-runtime to codeowners

0f91bb3

	// Drain all messages from the channel to remove messages sent to release the reference on
	// Drain all messages from the channel to remove messages sent to release the reference on IoStack

	.map_err(\|e\| anyhow::anyhow!(e.to_string()))?;
	.map_err(\|e\| anyhow::anyhow!(e))?;

	/// Since this method can return a single-threaded runtime it should only be use to
	/// Since this method can return a single-threaded runtime it should only be used to

		/// This function should not take ownership of the SharedRuntime as it will cause the runtime
		/// to be dropped in a non-blocking context causing a panic.

Conversation

VianneyRuhlmann commented Feb 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does this PR do?

Implementing the shared runtime

Update to the Worker trait

Expose ffi

Porting the existing workers

Stats exporter

Info fetcher

Telemetry worker

Losing telemetry payloads on shutdown

Using the shared runtime in TraceExporter

Motivation

Additional Notes

How to test the change?

Uh oh!

pr-commenter bot commented Feb 18, 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

scenario:benching deserializing traces from msgpack to their internal representation

scenario:concentrator/add_spans_to_concentrator

scenario:credit_card/is_card_number/ 3782-8224-6310-005

scenario:credit_card/is_card_number/ 378282246310005

scenario:credit_card/is_card_number/378282246310005

scenario:credit_card/is_card_number/37828224631000521389798

scenario:credit_card/is_card_number/x371413321323331

scenario:credit_card/is_card_number_no_luhn/ 378282246310005

scenario:credit_card/is_card_number_no_luhn/378282246310005

scenario:credit_card/is_card_number_no_luhn/37828224631000521389798

scenario:credit_card/is_card_number_no_luhn/x371413321323331

scenario:normalization/normalize_name/normalize_name/Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Long-.Too-Lo...

scenario:normalization/normalize_name/normalize_name/good

scenario:write only interface

Candidate

Group 1

Group 2

Group 3

Group 4

Group 5

Group 6

Group 7

Group 8

Group 9

Group 10

Group 11

Group 12

Group 13

Group 14

Group 15

Group 16

Group 17

Group 18

Group 19

Baseline

Uh oh!

github-actions bot commented Feb 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

📚 Documentation Check Results

📦 libdd-common - 168 warning(s)

📦 libdd-data-pipeline - 799 warning(s)

📦 libdd-telemetry - 478 warning(s)

Uh oh!

github-actions bot commented Feb 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Clippy Allow Annotation Report

Summary by Rule

Annotation Counts by File

Annotation Stats by Crate

About This Report

Uh oh!

github-actions bot commented Feb 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔒 Cargo Deny Results

📦 libdd-common - ✅ No issues

📦 libdd-data-pipeline - 1 error(s)

📦 libdd-telemetry - ✅ No issues

VianneyRuhlmann commented Feb 18, 2026 •

edited

Loading

pr-commenter bot commented Feb 18, 2026 •

edited

Loading

github-actions bot commented Feb 20, 2026 •

edited

Loading

📦 `libdd-common` - 168 warning(s)

📦 `libdd-data-pipeline` - 799 warning(s)

📦 `libdd-telemetry` - 478 warning(s)

github-actions bot commented Feb 20, 2026 •

edited

Loading

github-actions bot commented Feb 20, 2026 •

edited

Loading

📦 `libdd-common` - ✅ No issues

📦 `libdd-data-pipeline` - 1 error(s)

📦 `libdd-telemetry` - ✅ No issues