Add AgentOps Integration to LiteLLM (#9685 )

* feat(sidebars): add new item for agentops integration in Logging & Observability category

* Update agentops_integration.md to enhance title formatting and remove redundant section

* Enhance AgentOps integration in documentation and codebase by removing LiteLLMCallbackHandler references, adding environment variable configurations, and updating logging initialization for AgentOps support.

* Update AgentOps integration documentation to include instructions for obtaining API keys and clarify environment variable setup.

* Add unit tests for AgentOps integration and improve error handling in token fetching

* Add unit tests for AgentOps configuration and token fetching functionality

* Corrected agentops test directory

* Linting fix

* chore: add OpenTelemetry dependencies to pyproject.toml

* chore: update OpenTelemetry dependencies and add new packages in pyproject.toml and poetry.lock

2025-04-22 10:29:01 -07:00

2.2 KiB

Raw Blame History

🖇️ AgentOps - LLM Observability Platform

:::tip

This is community maintained. Please make an issue if you run into a bug: https://github.com/BerriAI/litellm

:::

AgentOps is an observability platform that enables tracing and monitoring of LLM calls, providing detailed insights into your AI operations.

Using AgentOps with LiteLLM

LiteLLM provides success_callbacks and failure_callbacks, allowing you to easily integrate AgentOps for comprehensive tracing and monitoring of your LLM operations.

Integration

Use just a few lines of code to instantly trace your responses across all providers with AgentOps: Get your AgentOps API Keys from https://app.agentops.ai/

import litellm

# Configure LiteLLM to use AgentOps
litellm.success_callback = ["agentops"]

# Make your LLM calls as usual
response = litellm.completion(
    model="gpt-3.5-turbo",
    messages=[{"role": "user", "content": "Hello, how are you?"}],
)

Complete Code:

import os
from litellm import completion

# Set env variables
os.environ["OPENAI_API_KEY"] = "your-openai-key"
os.environ["AGENTOPS_API_KEY"] = "your-agentops-api-key"

# Configure LiteLLM to use AgentOps
litellm.success_callback = ["agentops"]

# OpenAI call
response = completion(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hi 👋 - I'm OpenAI"}],
)

print(response)

Configuration Options

The AgentOps integration can be configured through environment variables:

AGENTOPS_API_KEY (str, optional): Your AgentOps API key
AGENTOPS_ENVIRONMENT (str, optional): Deployment environment (defaults to "production")
AGENTOPS_SERVICE_NAME (str, optional): Service name for tracing (defaults to "agentops")

Advanced Usage

You can configure additional settings through environment variables:

import os

# Configure AgentOps settings
os.environ["AGENTOPS_API_KEY"] = "your-agentops-api-key"
os.environ["AGENTOPS_ENVIRONMENT"] = "staging"
os.environ["AGENTOPS_SERVICE_NAME"] = "my-service"

# Enable AgentOps tracing
litellm.success_callback = ["agentops"]

Support

For issues or questions, please refer to:

2.2 KiB Raw Blame History