Skip to main content

Quick Start

Agenta enables you to capture all inputs, outputs, and metadata from your Large Language Model (LLM) applications, whether they're hosted within Agenta or running in your own environment. This guide will walk you through setting up observability for an OpenAI application running locally.

Why Instrument Your Application?​

Instrumenting your application allows you to monitor and debug its behavior in detail. You'll gain insights into how your app processes requests, helping you identify issues and optimize performance.

Key Concepts​

Traces: A trace represents the complete journey of a request through your application. In our context, a trace corresponds to a single request to your LLM application.

Spans: A span is a unit of work within a trace. Spans can be nested, forming a tree-like structure. The root span represents the overall operation, while child spans represent sub-operations. Agenta enriches each span with cost information and metadata when you make LLM calls.

Note: If you create an application through the Agenta UI, tracing is enabled by default. No additional setup is required—simply go to the observability view to see all your requests.

Step-by-Step Guide​

1. Install Required Packages​

First, install the Agenta SDK, OpenAI, and the OpenTelemetry instrumentor for OpenAI:

pip install -U agenta openai opentelemetry-instrumentation-openai

2. Obtain an API Key (Cloud and Enterprise Editions Only)​

If you're using Agenta Cloud or Enterprise Edition, you'll need an API key:

  1. Visit the Agenta API Keys page.
  2. Click on Create New API Key and follow the prompts.

3. Configure Environment Variables​

Set up the environment variables for your API key and host.

For Cloud and Enterprise Editions:

import os

os.environ["AGENTA_API_KEY"] = "YOUR_API_KEY_HERE"
os.environ["AGENTA_HOST"] = "https://cloud.agenta.ai"

For Open Source Version (OSS) Running Locally:

import os

# No API key is needed for OSS
os.environ["AGENTA_HOST"] = "http://localhost"

4. Instrument Your Application​

Below is a sample script to instrument an OpenAI application:

import agenta as ag
from opentelemetry.instrumentation.openai import OpenAIInstrumentor
import openai

# Initialize Agenta
ag.init()

# Instrument OpenAI
OpenAIInstrumentor().instrument()

# Set your OpenAI API key
openai.api_key = "YOUR_OPENAI_API_KEY"

# Make a chat completion request
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Write a short story about AI Engineering."},
    ],
)

# Print the assistant's reply
print(response.choices[0].message.content)

Explanation:

  • Import Libraries: Import Agenta, OpenAI, and the OpenTelemetry instrumentor.
  • Initialize Agenta: Call ag.init() to initialize the Agenta SDK.
  • Instrument OpenAI: Use OpenAIInstrumentor().instrument() to enable tracing for OpenAI calls.
  • Set OpenAI API Key: Replace "YOUR_OPENAI_API_KEY" with your actual OpenAI API key.
  • Make a Request: Send a chat completion request to the OpenAI API.
  • View Output: The assistant's response is printed to the console.

5. View Traces in the Agenta UI​

After running your application, you can view the captured traces in Agenta:

  1. Log in to your Agenta dashboard.
  2. Navigate to the Observability section.
  3. You'll see a list of traces corresponding to your application's requests.

Note: Replace path_to_screenshot_image with the actual path to your screenshot.

Next Steps​

Now that you've instrumented a simple OpenAI call, you can:

  • Instrument Specific Applications: Specify application IDs to monitor particular apps within Agenta.
  • Instrument Complex Workflows: Use decorators and instrument functions in more intricate workflows.
  • Add Metadata: Enrich your spans with custom metadata for deeper insights.
  • Explore Integrations: Auto-instrument other libraries and frameworks supported by Agenta:
    • OpenAI
    • LiteLLM
    • Anthropic
    • Bedrock
    • LangChain
    • LLamaIndex
    • Instructor
    • Vercel AI 0`