feat(profiling): Add JFR to OTLP/P converter with CLI, benchmarks and validation

[jbachorik]

What Does This Do

Adds the dd-java-agent/agent-profiling/profiling-otel module. It converts JFR recordings to OTLP Profiles protobufs (OTLP/P).

• Dictionary tables for strings, functions, locations, stacks, and links, with deduplication.
• A custom protobuf encoder with pinned OTLP proto field numbers (OtlpProtoFields).
• Converts jdk.ExecutionSample, jdk.NativeMethodSample, jdk.JavaMonitorEnter, jdk.JavaMonitorWait, jdk.ObjectAllocationSample, MethodSample, and jdk.CPUTimeSample to OTLP CPU samples.
• Resource attribute sharing and deduplication.
• A standalone CLI (JfrToOtlpConverterCLI, convert-jfr.sh).
• A JMH benchmark suite: dictionary tables, protobuf encoder, stack trace conversion, and full file conversion.
• A validation harness: OtlpProfileValidator with profcheck, pinned OTLP proto schemas, and an OTel collector test container. Includes docker/Dockerfile.profcheck.
• Module docs: ARCHITECTURE, BENCHMARKS, CLI, INTEGRATION_PLAN, PROFCHECK_INTEGRATION.
• Moves io.btrace:jafar to 0.26.2 and reworks the agent shadow jar bundling for jafar.

Motivation

The agent records JVM profiles as JFR recordings. An OTLP backend reads profiles only in the OTLP Profiles format. The agent must convert JFR to OTLP/P before it sends. This PR is that converter, as one module with no changes to existing code.

The jafar changes belong to this PR, because this module needs them. The converter reads JFR files through the jafar parser API (io.jafar.parser.api). Its tests write synthetic JFR recordings through org.openjdk.jmc:flightrecorder.writer. The version catalog gains the jafar-parser and jmc-flightrecorder-writer aliases for these two artifacts. jafar-tools is a fat jar that bundles an incomplete copy of the jafar-parser classes. The agent only needs the io.jafar.tools classes, and profiling-scrubber is their only consumer. The bundled parser classes carry no proper type metadata, and the GraalVM native-image builder fails on them. The new bundling drops the fat jar. The agent bundles jafar-parser through profiling-otel, and only the io.jafar.tools classes from the fat jar.

The module also carries a CLI, benchmarks, and a validation harness:

• The CLI converts recordings outside the agent. Developers use it to inspect converter output and to reproduce converter defects without an agent run.
• The benchmark suite measures the conversion cost. The conversion runs in the agent upload path, so its CPU and allocation cost must be known first.
• The validation harness checks the output twice: against pinned OTLP proto schemas (profcheck), and against a real OTel collector in a test container. This checks the wire format against a real receiver, not only against our encoder.

Additional Notes

Part of a 5-PR stack. Each PR bases on the previous one.

1. The payload and request body move (refactor: Move OtlpPayload and OTLP request bodies to communication/otlp-exporter #12880)
2. The sender extraction and the canonical resource attributes (refactor: Extract OTLP sender infrastructure to communication/otlp-exporter #12868)
3. RecordingData reference counting (feat(profiling): Reference-count RecordingData lifecycle #12869)
4. this PR: the JFR to OTLP/P converter, profiling-otel, plus jafar 0.26.2
5. OTLP profiles export wiring (feat(profiling): Integrate OTLP profiles export into the agent #12871)

Contributor Checklist

• Format the title according to the contribution guidelines
• Assign the type: and (comp: or inst:) labels in addition to any other useful labels
• Update the CODEOWNERS file on source file addition, migration, or deletion
• Update public documentation with any new configuration flags or behaviors
• Once approved, use merge queue to merge the PR

Jira ticket: [PROJ-IDENT]