Add support for agent identity propagation for compiled agents in nested graph

[Radhika Gupta (rads-1996)]

No description provided.

[github-actions]

Performance comparison

Threshold: regressions >15.0% on gating scenarios fail the build. Higher ops/s is better; positive Δ means the PR is slower.

Scenario	Gating	Baseline (ops/s)	Candidate (ops/s)	Δ %	Status
azure_monitor_log	yes	28,134.9	27,736.2	+1.44%	✅
azure_monitor_span	yes	159,261.0	159,235.7	+0.02%	✅
otel_log	no	32,682.9	32,334.2	+1.08%	✅
otel_span	no	47,980.0	47,773.7	+0.43%	✅

[Copilot]

Pull request overview

Adds auto-detection of nested compiled LangGraph subgraphs (“compiled agents”) so agent identity propagation uses the compiled graph’s own name (instead of the outer wrapper node name) when emitting spans.

Changes:

• Introduces _is_subgraph_boundary(run) to detect nested compiled-graph boundaries via run.name != langgraph_node (excluding generic names).
• Updates LangGraph node suppression and agent name resolution to keep/emphasize nested compiled-agent boundaries.
• Adds targeted unit tests covering boundary detection, suppression behavior, and end-to-end nested span emission/naming.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
tests/langchain/test_tracer.py	Adds helper + tests for subgraph-boundary detection, name resolution, and nested span emission/suppression.
src/microsoft/opentelemetry/_genai/_langchain/_tracer.py	Implements subgraph-boundary detection and integrates it into suppression and agent name resolution.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.