Quickstart Setup

Navigate to the app directory and activate the virtual environment you created in Monitoring Agentic AI Applications (or create a fresh one):

cd ~/workshop/agentic-ai/base-app
python3 -m venv .venv
source .venv/bin/activate

Install the Galileo SDK alongside the app’s existing dependencies in the requirements.txt:

pip install -r requirements.txt
pip install galileo python-dotenv

The app already installs langchain, langchain-openai, langgraph, and flask via its requirements.txt. The Galileo LangChain callback ships with the galileo package.

Your EC2 instance comes pre-configured with OPENAI_API_KEY and OPENAI_BASE_URL environment variables, which will be used by the application.

Add your credentials to the app’s .env file.

OPENAI_API_KEY="your-openai-api-key"
OPENAI_BASE_URL="https://lite-llm-proxy.splunko11y.com/v1"
GALILEO_API_KEY="your-galileo-api-key"
GALILEO_CONSOLE_URL="https://console.multitenant.galileocloud.io"
# Recommended: uncomment to group this workshop's traces under their own project
# and log stream. If you leave these commented out, Splunk Agent Observability uses a project and log
# stream both named "default".
# GALILEO_PROJECT="Workshop19"
# GALILEO_LOG_STREAM="TravelPlanner"

Uncommenting GALILEO_PROJECT and GALILEO_LOG_STREAM keeps the workshop traces easy to find. Leaving them commented is fine too — your traces will just land in the default project and default log stream.

4. Initialize Galileo near the top of main.py, right after the existing imports and load_dotenv() call. Passing the environment variables through means the project and log stream come from your .env when set, and fall back to Splunk Agent Observability’s default/default when not:

from dotenv import load_dotenv
load_dotenv()

import os
from galileo import galileo_context

galileo_context.init(project=os.getenv("GALILEO_PROJECT"),
                     log_stream=os.getenv("GALILEO_LOG_STREAM"))