Verified the AgentCreation type and TelemetryHandler.create_agent() context manager end-to-end by deploying a LanggraphAgent to Vertex AI Agent Engine.

Demo PR: #4336

Steps exercised:

• LLM call via google-genai SDK (auto-instrumented by VertexAIInstrumentor)
• Local agent creation wrapped in create_agent span
• Remote deploy to Agent Engine wrapped in create_agent span (~4 min)
• Query the deployed agent (1 USD = 9.2841 SEK)
• Cleanup (delete agent)

Trace 1:
create_agent Currency Exchange Agent (192µs, Client)
├── gen_ai.agent.description: Local LanggraphAgent for currency exchange
├── gen_ai.agent.name: Currency Exchange Agent
├── gen_ai.operation.name: create_agent
├── gen_ai.provider.name: gcp_vertex_ai
├── gen_ai.request.model: gemini-2.5-flash
└── server.address: us-central1-aiplatform.googleapis.com

Trace 2:
create_agent Currency Exchange Agent (4m5s, Client)
├── gen_ai.agent.description: Deploying agent to Vertex AI Agent Engine
├── gen_ai.agent.name: Currency Exchange Agent
├── gen_ai.operation.name: create_agent
├── gen_ai.provider.name: gcp_vertex_ai
├── gen_ai.request.model: gemini-2.5-flash
└── server.address: us-central1-aiplatform.googleapis.com

@etserend

Are you still working on this? Looks like there were some additions to genai-utils that you might have to rebase on.

@etserend

Are you still working on this? Looks like there were some additions to genai-utils that you might have to rebase on.

@lzchen Thanks for checking in. Per the SIG meeting, we're holding off on this implementation for now. I'll update the PR when we're ready to move forward.

thanks for the contribution @etserend !

We're moving GenAI instrumentations to the new repository - https://github.com/open-telemetry/opentelemetry-python-genai.

I'm closing this PR here, but you're welcome to reopen it in the new repo.