Gemini

Coralogix's AI Observability integrations provide deep visibility into applications that rely on Google Gemini models. With a dedicated integration for the Google Generative AI SDK, Coralogix delivers consolidated insight into synchronous, streaming, and async calls, enabling teams to monitor performance, cost, and reliability across every Gemini workload.

Overview

This library supplies production-ready OpenTelemetry instrumentation for the Google GenAI SDK. It automatically traces invocations of Gemini text and tool-calling workflows, enriches spans with model metadata, and records usage metrics. The integration accelerates debugging, highlights token consumption, and standardizes observability across Gemini-powered features.

Requirements

• Python version 3.9 and above.
• Coralogix API keys.
• google-genai package version 1.0.0 (minimum) and below 2.0.0.

Configure the Gemini SDK with a valid API key before activating the instrumentation.

Installation

Run the following command.

pip install llm-tracekit[gemini]

Authentication

Exporting spans to Coralogix requires configuring OTLP credentials before enabling instrumentation. Use setup_export_to_coralogix or the corresponding environment variables to supply authentication details.

Using setup_export_to_coralogix

from llm_tracekit import setup_export_to_coralogix

setup_export_to_coralogix(
    coralogix_token="<your_coralogix_token>",
    coralogix_endpoint="<your_coralogix_endpoint>",
    application_name="<ai-application>",
    subsystem_name="<ai-subsystem>",
    capture_content=True
)

Using environment variables

If arguments are not passed to setup_export_to_coralogix, the helper reads the following environment variables:

• CX_TOKEN: Your Coralogix API key.
• CX_ENDPOINT: ingress.:443, the OpenTelemetry endpoint associated with your Coralogix domain.
• CX_APPLICATION_NAME: Your application's name.
• CX_SUBSYSTEM_NAME: Your subsystem's name.

Usage

Set up instrumentation for the Google GenAI SDK.

Set up tracing

Instrument

To instrument all Gemini models, create an instance of GeminiInstrumentor and call the instrument method.

from llm_tracekit import GeminiInstrumentor

instrumentor = GeminiInstrumentor()
instrumentor.instrument()

Uninstrument

To remove instrumentation, call the uninstrument method.

instrumentor.uninstrument()

Full example

from google import genai

from llm_tracekit import setup_export_to_coralogix
from llm_tracekit import GeminiInstrumentor

setup_export_to_coralogix(
    coralogix_token="<your_coralogix_token>",
    coralogix_endpoint="<your_coralogix_endpoint>",
    application_name="<ai-application>",
    subsystem_name="<ai-subsystem>"
)

instrumentor = GeminiInstrumentor()
instrumentor.instrument()

client = genai.Client()

try:
    response = client.models.generate_content(
        model="gemini-2.0-flash-lite",
        contents="What's the temperature in Tel Aviv?"
    )
finally:
    client.close()

print(response)

Enable message content capture

By default, message content - such as prompts, completions, tool call arguments, and tool responses - is not captured.

To capture message content as span attributes, set the environment variable OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT to true.

Many Coralogix AI evaluations rely on message content, so enabling capture is strongly recommended.

Key differences from OpenTelemetry

• User prompts, tool invocations, and model responses are stored as span attributes instead of log events, preserving a single correlated timeline for each request.

Semantic conventions