LangGraph Monitoring (with LiteLLM)

This example routes LLM calls through a local LiteLLM proxy and logs prompts/outputs from each node.

Install

pip install langgraph langchain openai

Minimal Graph

from langgraph.graph import StateGraph, START, END
from typing import TypedDict
from openai import OpenAI

client = OpenAI(base_url="http://127.0.0.1:4000", api_key="sk-local")

class S(TypedDict):
    question: str
    answer: str

def llm_node(state: S):
    prompt = f"Answer briefly: {state['question']}"
    print("[prompt]", prompt)
    resp = client.chat.completions.create(
        model="ollama/llama3:8b",
        messages=[{"role": "user", "content": prompt}],
    )
    out = resp.choices[0].message.content
    print("[output]", out)
    return {"answer": out}

g = StateGraph(S)
g.add_node("llm", llm_node)
g.add_edge(START, "llm")
g.add_edge("llm", END)
app = g.compile()

print(app.invoke({"question": "What is LangGraph?"}))

Monitoring Options

• Console printing as shown above for quick visibility.
• Add structured logging or tracing hooks (e.g., to files or a DB).
• Optional: integrate with LangSmith to capture traces and artifacts.