Observability for Gemini with Datadog

OpenLLMetry lets you trace models running on VertexAI. With 5 minutes of work you can get complete view of your system directly into Datadog. See how below.

Step 1

Install the Traceloop SDK and initialize it. It will automatically log all calls to Gemini or any other VertexAI models, with prompts and completions as separate spans.

import google.generativeai as genai
from traceloop.sdk import Traceloop
from traceloop.sdk.decorators import workflow

Traceloop.init(app_name="joke_generation_service")

genai.configure(api_key=os.environ.get("GENAI_API_KEY"))

@workflow(name="joke_creation")
def create_joke():
    model = genai.GenerativeModel("gemini-1.0-pro-latest")
    response = model.generate_content(
        "Give me ten interview questions for the role of program manager.",
    )

    return response.text

Step 2

Configure your Datadog agent to receive traces from OpenLLMetry

# In your values.yaml
otlp:
  receiver:
    protocols:
      http:
        enabled: true
        
# In the environment of your application
TRACELOOP_BASE_URL=http://datadog-agent-hostname:4318

Docs

Observability for Gemini with Datadog

Step 1

Step 2

Discover use cases

More things you can connect to {platform}

Mistral AI

Chroma

Anthropic

together.ai

Ollama

LlamaIndex

OpenAI

Watsonx

Cohere

LangChain

Bedrock