[dougqh]

What Does This Do

Registers the CI Visibility tags in the tag registry, on span types for test sessions, modules, suites and tests, with requirement levels that tell the slot coloring which tags are set only some of the time. With #12893, every tag constant is now registered or deliberately excluded.

• Span types from the code. CI Visibility creates one span per test session, module, suite and test. Their span.kind values (test_session_end, test_module_end, test_suite_end, test) have no direction. Each tag's production set sites were checked for which of the four it lands on.
  • ci.span (abstract): what standard enrichment (TestDecoratorImpl.afterStart) puts on every CI span: the session name, test type, status, framework, session id, and the CI provider and commit (ci.*, git.*).
  • ci.aggregate (abstract, sessions and modules): command, coverage, ITR skipping counts, early flake detection, test management.
  • test.session, test.module, test.suite, test, sharing tags across levels by ref (test.module, test.suite, the ids, the source location). test.browser (Selenium) extends test.
  • Java overlay: test.gradle.session, test.gradle.module, test.maven.module, test.android (Robolectric), test.junit5, and junit.callback for JUnit's before and after spans.
• Requirement levels decide the slots. A test span carries about 100 tags, more than the dense store's 63 presence bits. With Color known tags into co-occurrence slots in their ids (phase 3a - groundwork - dense tag storage) #12882, a span type that overflows shares slots among its optional tags only, so the levels say which tags matter: what standard enrichment always sets (the CI provider and commit, ci.* / git.*) and the test source range are recommended, and the tags set only some of the time stay optional: pull request and head commit info, retries, ITR, test management, configuration errors, coverage, browser and Android details.

Motivation

The id migration needs every tag registered, and the dense store gives a tag a precise presence bit only when it's placed on the span types it appears on.

Falling back to "unslotted" for the overflow would have been the easy way out, but every slotless tag shares bit 63: set on most spans, so checking any of them means a scan. A designed collision between tags that are rarely present together keeps the fast miss for the tags that matter, and deriving it from requirement levels keeps slots out of the conventions.

Additional Notes

• Coloring: Color known tags into co-occurrence slots in their ids (phase 3a - groundwork - dense tag storage) #12882's coloring is now on master, and this branch has it. The 319 tags color into 63 slots: only the ten CI test span types (73 to 104 tags) overflow, and there 66 optional tags share slots: CI Visibility's occasional tags plus four optional tags inherited from base (_dd.span_links, _dd.svc_src, error.debug_info_captured, events). Every APM span type stays exact. The generator's output matches the earlier simulation exactly.
• Alternatives considered: a slot: marker in the conventions (an earlier revision; slots are the generator's business, and the level already carries the information), two occupancy longs (8 bytes and a word select on every span, for CI's sake), or a group bit like trace-level (doesn't help: the overflow is within one span's tags).
• The CI environment is stamped on every span. afterStart copies the ci.* / git.* block onto every session, module, suite and test span, though it's constant for the process. Setting it once per trace would be a CI Visibility behavior change (the intake may expect it on each event), for the CI Visibility team to decide; not changed here.
• Ids hoisted by the intake mapper are registered. test_session_id, test_module_id, test_suite_id and itr_correlation_id are set as tags and moved to top-level fields by CiTestCycleMapperV1, so they go through setTag and need ids. The os.* / runtime.* fields are tags only when CI Visibility reports through the APM protocol, so they're shared.
• Built on Register APM tag constants on their span types (otlp - tag registry) #12893, now merged. The first commit's slot: shared marker is reverted by the later two.

🤖 Generated with Claude Code

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - nebula-release-plugin

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	36.78	37.15	$\color{green}{\blacktriangledown}$ -0.37	37.15	$\color{green}{\blacktriangledown}$ -0.37	89/222
agentless	38.66	37.15	$\color{red}{\blacktriangle}$ +1.51	37.15	$\color{red}{\blacktriangle}$ +1.51	89/195
agentlessCodeCoverage	45.84	46.30	$\color{green}{\blacktriangledown}$ -0.46	45.38	$\color{red}{\blacktriangle}$ +0.46	89/194
agentlessLineCoverage	57.10	56.55	$\color{red}{\blacktriangle}$ +0.55	56.55	$\color{red}{\blacktriangle}$ +0.55	89/194

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - reactive-streams-jvm

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	21.87	22.54	$\color{green}{\blacktriangledown}$ -0.67	22.54	$\color{green}{\blacktriangledown}$ -0.67	119/254
agentless	18.51	19.59	$\color{green}{\blacktriangledown}$ -1.08	19.59	$\color{green}{\blacktriangledown}$ -1.08	115/218
agentlessCodeCoverage	19.68	20.39	$\color{green}{\blacktriangledown}$ -0.71	20.39	$\color{green}{\blacktriangledown}$ -0.71	116/219
agentlessLineCoverage	26.41	27.52	$\color{green}{\blacktriangledown}$ -1.11	27.52	$\color{green}{\blacktriangledown}$ -1.11	79/154

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - pass4s

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	4.72	10.54	$\color{green}{\blacktriangledown}$ -5.82	10.13	$\color{green}{\blacktriangledown}$ -5.41	90/222
agentless	5.97	9.92	$\color{green}{\blacktriangledown}$ -3.95	9.92	$\color{green}{\blacktriangledown}$ -3.95	89/192
agentlessCodeCoverage	12.75	17.38	$\color{green}{\blacktriangledown}$ -4.63	17.73	$\color{green}{\blacktriangledown}$ -4.98	80/176

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - netflix-zuul

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	91.33	89.58	$\color{red}{\blacktriangle}$ +1.75	89.58	$\color{red}{\blacktriangle}$ +1.75	89/223
agentless	83.87	82.69	$\color{red}{\blacktriangle}$ +1.18	82.69	$\color{red}{\blacktriangle}$ +1.18	89/194
agentlessCodeCoverage	98.23	97.04	$\color{red}{\blacktriangle}$ +1.19	97.04	$\color{red}{\blacktriangle}$ +1.19	89/194
agentlessLineCoverage	115.22	113.88	$\color{red}{\blacktriangle}$ +1.34	113.88	$\color{red}{\blacktriangle}$ +1.34	89/194

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - heliboard

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	10.60	9.73	$\color{red}{\blacktriangle}$ +0.87	9.92	$\color{red}{\blacktriangle}$ +0.68	88/193

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[dd-octo-sts]

🟢 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	13.94 s	13.95 s	[-0.7%; +0.6%] (no difference)
startup:insecure-bank:tracing:Agent	12.88 s	12.90 s	[-0.8%; +0.6%] (no difference)
startup:petclinic:appsec:Agent	17.11 s	16.89 s	[-4.8%; +7.4%] (unstable)
startup:petclinic:iast:Agent	17.26 s	17.04 s	[-3.1%; +5.7%] (no difference)
startup:petclinic:profiling:Agent	16.78 s	17.29 s	[-7.4%; +1.5%] (no difference)
startup:petclinic:sca:Agent	17.47 s	17.40 s	[-0.6%; +1.5%] (no difference)
startup:petclinic:tracing:Agent	16.10 s	16.53 s	[-6.6%; +1.4%] (no difference)

Commit: ed3d91a4 · 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.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - sonar-kotlin

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	13.98	13.67	$\color{red}{\blacktriangle}$ +0.31	13.40	$\color{red}{\blacktriangle}$ +0.58	101/236
agentless	12.12	12.37	$\color{green}{\blacktriangledown}$ -0.25	12.37	$\color{green}{\blacktriangledown}$ -0.25	101/206
agentlessCodeCoverage	16.03	15.72	$\color{red}{\blacktriangle}$ +0.31	15.41	$\color{red}{\blacktriangle}$ +0.62	101/206
agentlessLineCoverage	18.27	17.38	$\color{red}{\blacktriangle}$ +0.89	17.73	$\color{red}{\blacktriangle}$ +0.54	100/205

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - jolokia

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	98.17	97.04	$\color{red}{\blacktriangle}$ +1.13	95.12	$\color{red}{\blacktriangle}$ +3.05	104/238
agentless	94.44	91.39	$\color{red}{\blacktriangle}$ +3.05	91.39	$\color{red}{\blacktriangle}$ +3.05	104/210
agentlessCodeCoverage	103.50	101.00	$\color{red}{\blacktriangle}$ +2.50	99.00	$\color{red}{\blacktriangle}$ +4.50	104/210
agentlessLineCoverage	104.30	103.04	$\color{red}{\blacktriangle}$ +1.26	101.00	$\color{red}{\blacktriangle}$ +3.30	115/227

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - okhttp

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	18.82	19.20	$\color{green}{\blacktriangledown}$ -0.38	19.59	$\color{green}{\blacktriangledown}$ -0.77	104/237
agentless	18.78	18.45	$\color{red}{\blacktriangle}$ +0.33	19.20	$\color{green}{\blacktriangledown}$ -0.42	103/212
agentlessCodeCoverage	23.53	21.65	$\color{red}{\blacktriangle}$ +1.88	22.54	$\color{red}{\blacktriangle}$ +0.99	105/216
agentlessLineCoverage	38.13	37.91	$\color{red}{\blacktriangle}$ +0.22	38.67	$\color{green}{\blacktriangledown}$ -0.54	102/210

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - spring_boot

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	15.93	16.36	$\color{green}{\blacktriangledown}$ -0.43	16.36	$\color{green}{\blacktriangledown}$ -0.43	101/234
agentless	9.25	9.92	$\color{green}{\blacktriangledown}$ -0.67	9.92	$\color{green}{\blacktriangledown}$ -0.67	101/206
agentlessCodeCoverage	13.51	13.94	$\color{green}{\blacktriangledown}$ -0.43	13.67	$\color{green}{\blacktriangledown}$ -0.16	101/208
agentlessLineCoverage	23.65	22.99	$\color{red}{\blacktriangle}$ +0.66	22.54	$\color{red}{\blacktriangle}$ +1.11	100/205

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[cit-pr-commenter-54b7da]

CI Visibility Test Environment - sonar-java

Job Status: 🟢 success

Scenario	This PR (%)	7d median	Δ 7d	30d median	Δ 30d	runs (7d/30d)
agent	-26.66	14.51	$\color{green}{\blacktriangledown}$ -41.17	13.94	$\color{green}{\blacktriangledown}$ -40.60	101/234
agentless	9.49	12.62	$\color{green}{\blacktriangledown}$ -3.13	12.87	$\color{green}{\blacktriangledown}$ -3.38	101/208
agentlessCodeCoverage	36.55	79.45	$\color{green}{\blacktriangledown}$ -42.90	79.45	$\color{green}{\blacktriangledown}$ -42.90	101/208
agentlessLineCoverage	58.72	120.92	$\color{green}{\blacktriangledown}$ -62.20	123.36	$\color{green}{\blacktriangledown}$ -64.64	100/208

Baseline: median of @test.tracer_overhead on main (gitlab) over the last 7/30 days, per OSS project & scenario. Δ = this PR − baseline median; red ▲ = more overhead, green ▽ = less overhead than baseline.

Tests run to verify that CI Visibility behavior has not regressed in the current PR.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-20T18:10:24.566757Z	ed3d91a	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ed3d91a464

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]

Model library_version on every span in CI traces

When a CI test produces ordinary child spans or JUnit callback spans, CiVisibilityTraceInterceptor.onTraceComplete adds library_version to every span in the trace, not only to session/module/suite/test events. Declaring it solely under ci.span therefore omits its co-occurrence with regular internal, HTTP, database, and other span tags from the coloring graph; the generated assignment can reuse its slot for a tag that is present on such a child span, defeating the exact presence bit and forcing dense-store lookups to scan. Model this as a CI-gated tag applying to all span types rather than as a ci.span-only tag.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Register the library capability tags

Whenever CI Visibility advertises library capabilities, TestImpl sets the finite _dd.library_capabilities.<capability> tags produced by LibraryCapability.asTag(), but none of those ten names is declared in either conventions file. Because the coverage test only reflects constants from Tags, InstrumentationTags, and DDTags, it still passes while KnownTagCodec.keyOf returns zero for these live test tags, leaving them on the custom-tag path and making the stated ID/dense-layout migration incomplete. Add the enum's capability keys as optional string tags on test and include this generated family in registry coverage.

Useful? React with 👍 / 👎.

[datadog-official]

More details

The registered CI IDs and APM-only OS/runtime tags match their setter usage. Registry generation, compilation, and direct tag-map checks supported compatibility.

Was this helpful? React 👍 or 👎

_{Open Bits AI session}

_{🤖 Bits Code Review · Commit ed3d91a}