GenAI spec updates for langchain

.gitignore

@@ -61,3 +61,4 @@ target
 
 # Benchmark result files
 *-benchmark.json
+instrumentation-genai/opentelemetry-instrumentation-langchain/examples/multi_agent_travel_planner/.env

instrumentation-genai/README.md

@@ -2,7 +2,7 @@
 | Instrumentation | Supported Packages | Metrics support | Semconv status |
 | --------------- | ------------------ | --------------- | -------------- |
 | [opentelemetry-instrumentation-google-genai](./opentelemetry-instrumentation-google-genai) | google-genai >= 1.0.0 | No | development
-| [opentelemetry-instrumentation-langchain](./opentelemetry-instrumentation-langchain) | langchain >= 0.3.21 | No | development
+| [opentelemetry-instrumentation-langchain](./opentelemetry-instrumentation-langchain) | langchain >= 1.0.0 | No | development
 | [opentelemetry-instrumentation-openai-agents-v2](./opentelemetry-instrumentation-openai-agents-v2) | openai-agents >= 0.3.3 | No | development
 | [opentelemetry-instrumentation-openai-v2](./opentelemetry-instrumentation-openai-v2) | openai >= 1.26.0 | Yes | development
 | [opentelemetry-instrumentation-vertexai](./opentelemetry-instrumentation-vertexai) | google-cloud-aiplatform >= 1.64 | No | development

instrumentation-genai/opentelemetry-instrumentation-langchain/CHANGELOG.md

@@ -8,4 +8,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## Unreleased
 
 - Added span support for genAI langchain llm invocation.
-  ([#3665](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3665))
+  ([#3665](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3665))
+- Align LangChain instrumentation with GenAI schema 1.37.0 and add unit coverage for updated metadata.
+  ([#3813](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3813))
+- Capture GenAI input/output messages on spans by default with opt-out control.
+  ([#3813](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3813))

instrumentation-genai/opentelemetry-instrumentation-langchain/examples/manual/main.py

@@ -21,7 +21,7 @@ def main():
 
     # ChatOpenAI
     llm = ChatOpenAI(
-        model="gpt-3.5-turbo",
+        model="gpt-4.1",
         temperature=0.1,
         max_tokens=100,
         top_p=0.9,

instrumentation-genai/opentelemetry-instrumentation-langchain/examples/manual/requirements.txt

@@ -1,7 +1,7 @@
-langchain==0.3.21
-langchain_openai
+langchain==1.0.1
+langchain-openai==1.0.0
 opentelemetry-sdk>=1.31.0
 opentelemetry-exporter-otlp-proto-grpc>=1.31.0
 
 # Uncomment after lanchain instrumetation is released
-# opentelemetry-instrumentation-langchain~=2.0b0.dev
+# opentelemetry-instrumentation-langchain~=2.0b0.dev

instrumentation-genai/opentelemetry-instrumentation-langchain/examples/multi_agent_travel_planner/.env.example

@@ -0,0 +1,14 @@
+# Copy this file to .env and update values before running the sample.
+
+# Required OpenAI API key
+OPENAI_API_KEY=sk-YOUR_API_KEY
+
+# Optional: override default model (defaults to gpt-4o-mini)
+# OPENAI_MODEL=gpt-4o-mini
+
+# OTLP exporter configuration (update for your collector)
+# OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317
+# OTEL_EXPORTER_OTLP_PROTOCOL=grpc
+
+# Traces will use this service.name
+OTEL_SERVICE_NAME=opentelemetry-python-langchain-multi-agent

instrumentation-genai/opentelemetry-instrumentation-langchain/examples/multi_agent_travel_planner/.env.sample

@@ -0,0 +1,14 @@
+# Copy this file to `.env` and update values before running the sample.
+
+# Required OpenAI API key
+OPENAI_API_KEY=sk-YOUR_API_KEY
+
+# Optional: override default model (defaults to gpt-4.1)
+# OPENAI_MODEL=gpt-4.1
+
+# OTLP exporter configuration (update for your collector)
+# OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317
+# OTEL_EXPORTER_OTLP_PROTOCOL=grpc
+
+# Traces will use this service.name
+OTEL_SERVICE_NAME=opentelemetry-python-langchain-multi-agent

instrumentation-genai/opentelemetry-instrumentation-langchain/examples/multi_agent_travel_planner/README.rst

@@ -0,0 +1,62 @@
+Multi-Agent Travel Planner Sample
+=================================
+
+This example shows how to orchestrate a small team of LangChain agents with
+`LangGraph <https://python.langchain.com/docs/langgraph/>`_ while the
+OpenTelemetry LangChain instrumentation captures GenAI spans and forwards them
+to an OTLP collector.
+
+The graph contains four specialists (coordinator, flights, hotels, activities)
+and a final synthesiser node that produces an itinerary.  Each specialist relies
+on a simple, deterministic tool so you can run the example without any external
+travel APIs while still observing tool spans wired up to the agent calls.
+
+Prerequisites
+-------------
+
+* Python 3.10+
+* An OpenAI API key with access to ``gpt-4o-mini`` (or set ``OPENAI_MODEL`` to a
+  model that is available to your account)
+* A running OTLP collector (gRPC on ``localhost:4317`` by default)
+
+Setup
+-----
+
+.. code-block:: bash
+
+   cd instrumentation-genai/opentelemetry-instrumentation-langchain/examples/multi_agent_travel_planner
+   python3 -m venv .venv
+   source .venv/bin/activate
+   pip install -r requirements.txt
+
+   # Copy the sample environment file and update values as needed.
+   cp .env.example .env
+   source .env
+
+Run the sample
+--------------
+
+.. code-block:: bash
+
+   # From this directory, after activating the virtual environment and sourcing
+   # your environment variables:
+   python main.py
+
+The script prints each agent's contribution followed by the final itinerary.
+At the same time it streams OTLP traces.  You should see:
+
+* A root span named ``invoke_agent travel_multi_agent_planner`` that captures
+  the overall orchestration, including ``gen_ai.input.messages`` and a preview
+  of the final plan.
+* LangChain instrumentation spans for each agent's LLM invocation with
+  ``gen_ai.provider.name=openai`` and ``service.name`` derived from
+  ``OTEL_SERVICE_NAME``.
+
+Tear down
+---------
+
+Deactivate the virtual environment when you are done:
+
+.. code-block:: bash
+
+   deactivate