✨ อัปเดตล่าสุด: DataDog/dd trace java/pull/

MilanGarnier · 2026-09-03T13:05:58Z

What Does This Do

Implements OpenTelemetry consistent probability sampling with the ot
tracestate member.

Records locally made probability decisions as rv and th for agent-rate
and rule-based sampling (and forwards them is they were extracted and valid).
Keeps rate-limiter outcomes and manual overrides from creating contradictory
probability state.
Validates and normalizes inbound ot fields, rebuilding W3C tracestate
within its member and size limits while preserving unmanaged vendor members.
Publishes sampling priority, Datadog sampling tags, and OTel sampling state
together so W3C injection and OTLP JSON/protobuf export use one decision.
Clears inherited managed OTel state when Org Guard resamples a trace.

Adds coverage for parsing and normalization, W3C extraction and
reinjection, sampling decisions and overrides, Org Guard, and OTLP export.

Motivation

OpenTelemetry consistent probability sampling lets tracers make compatible
sampling decisions using a shared random value and threshold. The Java tracer
needs to retain compatible inbound state and publish locally made probability
decisions without conflicting with Datadog sampling behavior.

Additional Notes

Builds on the ot tracestate parser added in Parse OpenTelemetry tracestate member #12405, which has merged.
Follow-up system-tests coverage is being enabled in
DataDog/system-tests#7649.

Contributor Checklist

Format the title according to the contribution guidelines.
Assign the required type: and comp: labels.

Jira ticket: APMAPI-2171

datadog-datadog-prod-us1-2 · 2026-09-03T13:20:46Z

🎯 Code Coverage (details)
• Patch Coverage: 89.22%
• Overall Coverage: 59.24% (+0.14%)

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

🔗 Commit SHA: c2f6051 | Docs | Give us feedback!}

dd-octo-sts · 2026-09-03T13:27:22Z

🟢 Java Benchmark SLOs — All performance SLOs passed

Suite	Status
Startup	🟢 pass

SLO thresholds are defined here based on automatically generated metrics. A warning is raised when results are within 5% of the threshold.

PR vs. master results

Scenario	Candidate	master	Δ (95% CI of mean)
startup:insecure-bank:iast:Agent	14.79 s	14.70 s	[-0.3%; +1.5%] (no difference)
startup:insecure-bank:tracing:Agent	13.67 s	13.63 s	[-0.5%; +1.1%] (no difference)
startup:petclinic:appsec:Agent	17.59 s	17.36 s	[+0.4%; +2.3%] (maybe worse)
startup:petclinic:iast:Agent	16.72 s	17.61 s	[-9.2%; -0.8%] (maybe better)
startup:petclinic:profiling:Agent	17.25 s	17.25 s	[-1.2%; +1.3%] (no difference)
startup:petclinic:sca:Agent	17.70 s	17.53 s	[-0.1%; +2.0%] (no difference)
startup:petclinic:tracing:Agent	16.69 s	16.85 s	[-1.9%; +0.0%] (no difference)

Commit: c2f6051e · CI Pipeline · Benchmarking Platform UI

Load and DaCapo benchmarks can be triggered manually in the GitLab pipeline. Results will appear in the Benchmarking Platform UI after completion.

vandonr

ok for IDM

dd-octo-sts · 2026-09-29T09:55:29Z

/merge

gh-worker-devflow-routing-ef8351 · 2026-09-29T09:55:33Z

View all feedbacks in Devflow UI.

2026-09-29 09:55:33 UTC ℹ️ Start processing command /merge

2026-09-29 09:55:37 UTC ℹ️ MergeQueue: pull request added to the queue

The expected merge time in master is approximately 1h (p90).

2026-09-29 11:08:30 UTC ℹ️ MergeQueue: This merge request was merged

…hape cost SamplingDecisionBenchmark gains arms that follow a probability decision with the Datadog header, the W3C tracestate header and the OTLP tracestate, so the deferred OtelTraceState is measured where it's built or skipped. ServiceShapePropagationBenchmark measures head, leaf and intermediate services with a default-configured tracer. It only uses API unchanged since v1.66.0, so it can be built onto release tags to put a cost on #12397. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

…ported (#12982) * Defer building the OpenTelemetry tracestate for fresh probability decisions Every root probability decision built an OtelTraceState inside the sampling lock, though only W3C injection and OTLP export read it. Install a ProbabilitySamplingState that keeps the trace id bits and rate, and builds the ot member on first read. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Clarify why lazy otel state readers agree on one instance Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Add benchmarks for the lazy otel state's inject, export and service-shape cost SamplingDecisionBenchmark gains arms that follow a probability decision with the Datadog header, the W3C tracestate header and the OTLP tracestate, so the deferred OtelTraceState is measured where it's built or skipped. ServiceShapePropagationBenchmark measures head, leaf and intermediate services with a default-configured tracer. It only uses API unchanged since v1.66.0, so it can be built onto release tags to put a cost on #12397. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Rename ServiceShapePropagationBenchmark to PropagationScenariosBenchmark Describe it as a scenario benchmark: each arm is a whole request driven through the tracer's public API, rather than one method in isolation. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Test that racing readers get one OpenTelemetry state instance Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

…etry tracestate (#12981) Precompute decision maker tag values per sampling mechanism Every keep decision built "-" + samplingMechanism and then validated, hashed and looked it up in TagValue's cache. The mechanisms are a small fixed set, so build their TagValues once and index by mechanism, falling back to TagValue.from for anything beyond the table. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Add a sampling decision benchmark Measures the per-trace propagation-tag bookkeeping done on a root sampling decision, both in a bare start/sample/finish span and on fresh PropagationTags with no span around it, for the default, rule and manual keep paths. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Defer building the OpenTelemetry tracestate until it's injected or exported (#12982) * Defer building the OpenTelemetry tracestate for fresh probability decisions Every root probability decision built an OtelTraceState inside the sampling lock, though only W3C injection and OTLP export read it. Install a ProbabilitySamplingState that keeps the trace id bits and rate, and builds the ot member on first read. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Clarify why lazy otel state readers agree on one instance Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Add benchmarks for the lazy otel state's inject, export and service-shape cost SamplingDecisionBenchmark gains arms that follow a probability decision with the Datadog header, the W3C tracestate header and the OTLP tracestate, so the deferred OtelTraceState is measured where it's built or skipped. ServiceShapePropagationBenchmark measures head, leaf and intermediate services with a default-configured tracer. It only uses API unchanged since v1.66.0, so it can be built onto release tags to put a cost on #12397. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Rename ServiceShapePropagationBenchmark to PropagationScenariosBenchmark Describe it as a scenario benchmark: each arm is a whole request driven through the tracer's public API, rather than one method in isolation. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Test that racing readers get one OpenTelemetry state instance Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: devflow.devflow-routing-intake <devflow.devflow-routing-intake@kubernetes.us1.ddbuild.io>

MilanGarnier added type: feature Enhancements and improvements comp: context propagation Trace context propagation tag: ai generated Largely based on code generated by an AI or LLM labels Sep 3, 2026

MilanGarnier changed the title ~~Support OpenTelemetry tracestate sampling~~ feat(sampling): Support for OpenTelemetry consistent tracestate sampling Sep 3, 2026

MilanGarnier force-pushed the milan.garnier/ot.th branch from 3643548 to 2f90ba6 Compare September 3, 2026 13:58

MilanGarnier changed the base branch from master to milan.garnier/preserve-deterministic-sampler-rate September 3, 2026 13:59

Base automatically changed from milan.garnier/preserve-deterministic-sampler-rate to master September 3, 2026 18:45

MilanGarnier force-pushed the milan.garnier/ot.th branch from a168627 to 9e2422a Compare September 4, 2026 09:29

MilanGarnier mentioned this pull request Sep 4, 2026

Parse OpenTelemetry tracestate member #12405

Merged

3 tasks done

MilanGarnier changed the base branch from master to milan.garnier/otel-tracestate-refactor September 4, 2026 09:30

MilanGarnier mentioned this pull request Sep 4, 2026

test(java): enable ot.th/ot.rv tracestate sampling scenarios DataDog/system-tests#7649

Merged

MilanGarnier force-pushed the milan.garnier/ot.th branch 5 times, most recently from 7a1ddea to a0176c9 Compare September 8, 2026 14:26

This comment was marked as resolved.

Sign in to view

MilanGarnier marked this pull request as ready for review September 10, 2026 09:02

MilanGarnier requested review from a team as code owners September 10, 2026 09:02

MilanGarnier requested review from mhlidd and removed request for a team September 10, 2026 09:02

MilanGarnier force-pushed the milan.garnier/otel-tracestate-refactor branch from db321c0 to 1c02400 Compare September 10, 2026 09:02

MilanGarnier force-pushed the milan.garnier/ot.th branch from 3c81345 to 629892d Compare September 10, 2026 09:02

This comment was marked as resolved.

Sign in to view

MilanGarnier force-pushed the milan.garnier/ot.th branch from d5a6721 to eb4c156 Compare September 11, 2026 13:28

MilanGarnier and others added 7 commits September 28, 2026 18:26

fix(propagation): preserve opaque otel tracestate members

9125270

Merge remote-tracking branch 'origin/master' into milan.garnier/ot.th

367c960

perf: bound ot= unknown-field colon scan to member end

c5a8fea

perf: add identity short-circuit for forNonProbabilityDecision()

3a46a36

fix(propagation): clamp ot= threshold for out-of-range sampling rates

306576f

reuse LongStringUtils.parseUnsignedLongHex

8ce589d

reuse assertDefaultFallbackSampling

cdde406

mcculls force-pushed the milan.garnier/ot.th branch from ea1fb86 to cdde406 Compare September 28, 2026 22:24

mcculls approved these changes Sep 28, 2026

View reviewed changes

perf: avoid allocating temporary SamplingState on every extract

c2f6051

vandonr approved these changes Sep 29, 2026

View reviewed changes

MilanGarnier enabled auto-merge September 29, 2026 09:45

MilanGarnier disabled auto-merge September 29, 2026 09:45

MilanGarnier enabled auto-merge September 29, 2026 09:45

MilanGarnier added this pull request to the merge queue Sep 29, 2026

github-merge-queue Bot removed this pull request from the merge queue due to failed status checks Sep 29, 2026

gh-worker-dd-mergequeue-cf854d Bot merged commit 79a1710 into master Sep 29, 2026
606 checks passed

gh-worker-dd-mergequeue-cf854d Bot deleted the milan.garnier/ot.th branch September 29, 2026 11:08

github-actions Bot added this to the 1.67.0 milestone Sep 29, 2026

ValentinZakharov changed the title ~~feat(sampling): Support for OpenTelemetry consistent tracestate sampling~~ Support OpenTelemetry consistent tracestate sampling Oct 5, 2026

dougqh mentioned this pull request Oct 21, 2026

Defer parsing the extracted OpenTelemetry tracestate member #12987

Draft

dougqh mentioned this pull request Oct 23, 2026

Defer parsing the inherited OpenTelemetry tracestate member #13028

Draft

Navigation Menu

Repository navigation

Support OpenTelemetry consistent tracestate sampling - #12397

Support OpenTelemetry consistent tracestate sampling#12397
gh-worker-dd-mergequeue-cf854d[bot] merged 39 commits into
masterfrom
milan.garnier/ot.th

MilanGarnier commented Sep 3, 2026 •

edited

Loading

Uh oh!

datadog-datadog-prod-us1-2 Bot commented Sep 3, 2026 •

edited by datadog-datadog-prod-us1 Bot

Loading

Uh oh!

dd-octo-sts Bot commented Sep 3, 2026 •

edited

Loading

Uh oh!

This comment was marked as resolved.

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

vandonr left a comment

Uh oh!

dd-octo-sts Bot commented Sep 29, 2026

Uh oh!

gh-worker-devflow-routing-ef8351 Bot commented Sep 29, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Conversation

MilanGarnier commented Sep 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What Does This Do

Motivation

Additional Notes

Contributor Checklist

Uh oh!

datadog-datadog-prod-us1-2 Bot commented Sep 3, 2026 • edited by datadog-datadog-prod-us1 Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dd-octo-sts Bot commented Sep 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🟢 Java Benchmark SLOs — All performance SLOs passed

Uh oh!

This comment was marked as resolved.

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

vandonr left a comment

Choose a reason for hiding this comment

Uh oh!

dd-octo-sts Bot commented Sep 29, 2026

Uh oh!

gh-worker-devflow-routing-ef8351 Bot commented Sep 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

MilanGarnier commented Sep 3, 2026 •

edited

Loading

datadog-datadog-prod-us1-2 Bot commented Sep 3, 2026 •

edited by datadog-datadog-prod-us1 Bot

Loading

dd-octo-sts Bot commented Sep 3, 2026 •

edited

Loading

gh-worker-devflow-routing-ef8351 Bot commented Sep 29, 2026 •

edited

Loading