Monitoring Google Gemini models

Edit online

Google Gemini is a family of multimodal AI models designed to understand and process information across text, code, images, audio, and video. This guide shows you how to instrument an application using Google Gemini models with OpenLLMetry to send telemetry data to Instana.

Prerequisites

Make sure that the following prerequisites are met:

Python 3.8 or later
A Google API key for Gemini (get one from Google AI Studio)
An Instana backend account with Generative AI observability enabled
Review of Getting Started with Generative AI Observability on agent and agentless modes

Instrumenting your Google Gemini application

Install the required packages.
```
pip install google-genai traceloop-sdk
```

Export your Google API key.

export GOOGLE_API_KEY="<your-google-api-key>"

Create your Google Gemini application. Create a Python file with the following code:

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

# Initialize OpenLLMetry
Traceloop.init(app_name="gemini_chat_app", disable_batch=True)

# Initialize Gemini client
client = genai.Client()

@workflow(name="gemini_conversation")
def ask_gemini(question: str):
    """Send a question to Gemini and get a response."""

    chat = client.chats.create(model="gemini-2.0-flash-exp")
    response = chat.send_message(question)

    return response.text

# Example usage
if __name__ == "__main__":
    questions = [
        "What is Artificial Intelligence?",
        "How does machine learning work?"
    ]

    for question in questions:
        print(f"\nQuestion: {question}")
        answer = ask_gemini(question)
        print(f"Answer: {answer}\n")
        print("-" * 80)

Run your application.
```
python3 gemini_app.py
```
The application will send questions to Gemini and display the responses. OpenLLMetry automatically captures traces for each API call and sends them to Instana.
View data on Instana.

After running your application, the following items are displayed on the Instana Gen AI observability dashboard:
- Model used
- Token usage (input and output tokens)
- Response latency
- Request and response content

Troubleshooting

For common issues such as traces not appearing or connection errors, see Troubleshooting.

Authentication errors

If you encounter authentication errors:

Verify your GOOGLE_API_KEY is set correctly
Check whether your API key is valid in Google AI Studio
Make sure that your API key is not expired or revoked
Verify your API key has access to the Gemini API

Rate limiting errors

If you encounter rate limit errors:

Check your Google AI Studio account's rate limits
Add delays between requests if making multiple calls
Consider upgrading your plan for higher limits
Implement exponential backoff for retries

Model not found errors

If you encounter model not found errors:

Verify the model name is correct (for example, gemini-2.0-flash-exp, gemini-pro)
Check whether the model is available in your region
Make sure your API key has access to the specified model
Refer to Google's Gemini documentation for available models

Next steps

Explore LLM providers supported by Instana
Learn about cost calculation for your LLM usage
Set up alerts for your Google Gemini API usage
Review Google Gemini documentation for model capabilities