[dougqh]

What Does This Do

DDSpanContext owns its tag stores directly, instead of one TagMap:

• Known tags go in an IdStore, which is also the span's tag lock.
• Custom tags go in a HashStore, borrowed read-only from the read-through parent until the span's first custom write.

Each write is routed to the store that holds the tag. A name is resolved once and routed, and a write by id goes straight to the id store.

• The eager post-processors already read and write the span through SpanContents (Introduce SpanContents for eager tag post-processors (dense store - part 3) #13001), so they need no TagMap.
• The lazy post-processors and the serializer still take a TagMap. The span builds a view over its stores once, at serialization, and caches it. That costs about 32 B per span until they move to SpanContents.
• TagMap.Internals is gone, since the stores are now the direct path. The forbidden-API entry for it goes too.
• The constructors drop the unused tagsSize parameter, which is why many tests change by one line.

Motivation

This is the last of three PRs that replace the span's single TagMap with two stores split by tag kind. The stores and the TagMap facade are in the two PRs below this one. Here, the span stops going through TagMap for its own reads and writes. That gives the dense store (#13004) a place as the known-tag half.

Additional Notes

Allocation, 7-span web-request trace, application thread only:

step	base	this PR
bare spans	4,448 B	4,504 B
whole trace	7,936 B	8,040 B (+104 B, +1.3%)

• +56 B is the new DDSpanContext fields (customTags, customTagsBorrowed, tagView), 8 B per span.
• +48 B appears when the first custom tag is written. That's most likely the custom HashStore being created, but it isn't confirmed yet.
• Without SpanContents, building the view on the application thread made this +4.1%.

Tests: internal-api and dd-trace-core pass. The only failures are the two TracerConnectionReliabilityTest cases, which need Docker.

Where this is going

This PR is one step in replacing the span's single TagMap with storage that knows each tag's kind. The goal is to cut what every span allocates on the application thread, without changing behavior.

The end state

• A tag is classified once, at the boundary. A known tag becomes an id (from the tag registry, Resolve direction-scoped OpenTelemetry names at the OpenTelemetry boundary (otlp - tag registry - phase 3) #12731), and a custom tag stays a name. Instrumentation sets known tags by id (Route intercepted tags by id instead of by name (otlp - tag registry - phase 3) #12817), so the JIT can constant-fold the routing.
• The span owns its tag stores directly. Known tags are kept by id and custom tags by name. TagMap becomes a thin view for code outside the span.
• Known tags live in a dense store by default. Parallel arrays of serials and values replace an Entry object per tag, and each span's store is presized from a per-operation hint. With the flag on, a web-request-shaped trace allocates about 30% less on the application thread than on master.
• Core reads and writes a span through SpanContents, not a TagMap. That covers post-processors now and serializers later, so no TagMap view gets built for them. SpanContents starts with four methods and grows a name, id and primitive API as more readers move to it.

Later, not in this stack: cheaper read-through to parent spans, trace-level tags in a per-trace holder (TraceSegment), and primitive tag values without boxing.

The stack. This replaces #12918, split into pieces that can each be reviewed, and measured, on their own.

PR	Step	Allocation per 7-span trace (flag off unless noted)
#12999	A null tag value removes the tag in TagMap. This is the stack's only behavior change, kept separate on purpose.	unchanged
#12992	Bucket known tags by slot in TagMap, and add TraceShapeBenchmark	−96 B vs master
#13001	SpanContents for the eager post-processors	unchanged
#13002	Extract TagMap's bucket mechanics into HashSupport (a pure move)	unchanged
#13006	Add the id and name tag stores, not yet used	unchanged
#13007	Make TagMap a facade over the two stores	unchanged
#13003	Have the span own its tag stores directly	+104 B (+1.3%)
#13004	Dense id store behind trace.experimental.dense.tags.enabled (off by default)	flag on: −30.6%
#12926	Turn the dense store on by default, add sizing hints and bulk prototype seeding	flag on: about 3% less again

Every PR after #12999 is behavior-preserving.

The pieces' allocation figures come from a deterministic per-thread probe (ThreadMXBean allocated bytes) that builds the same 7-span trace as TraceShapeBenchmark. The #12992 and #12926 figures come from TraceShapeBenchmark itself.

🤖 Generated with Claude Code

[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	13.91 s	13.91 s	[-0.6%; +0.7%] (no difference)
startup:insecure-bank:tracing:Agent	12.90 s	12.90 s	[-0.7%; +0.6%] (no difference)
startup:petclinic:appsec:Agent	17.08 s	17.09 s	[-0.9%; +0.8%] (no difference)
startup:petclinic:iast:Agent	16.90 s	16.96 s	[-1.2%; +0.5%] (no difference)
startup:petclinic:profiling:Agent	16.72 s	16.16 s	[-0.9%; +7.8%] (no difference)
startup:petclinic:sca:Agent	17.07 s	17.04 s	[-0.7%; +1.1%] (no difference)
startup:petclinic:tracing:Agent	16.16 s	15.68 s	[-1.2%; +7.4%] (no difference)

Commit: 074d3dbe · 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.