The code we build in the Day 2 live session, kept deliberately small.

Day 1 turned DevOps functions into an API. Day 2 turns functions into tools an AI agent can call. The agent has just two tools:

1. Read a log file and count INFO / WARNING / ERROR lines.
2. List the Docker containers running on this machine (using subprocess).

The rule that keeps it reliable: the tool does the work, the LLM writes the story. The log function is plain Python and unit-tested without any LLM; the agent only decides which tool to call and summarizes the result.

• devops_tools.py — the two functions: count_log_levels and list_containers.
• agent.py — wraps them as tools and builds the agent with create_agent.
• test_tools.py — tests for the log function (no Ollama, no Docker needed).
• app.log — sample log (INFO=10, WARNING=2, ERROR=3).

Ollama running with a model pulled:

ollama serve
ollama pull llama3.2
curl http://localhost:11434/api/tags

Docker is only needed for the container tool. Without it, that tool returns a friendly "Docker is not installed" message instead of crashing.

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

python -m pytest .     # the log function (no Ollama needed)
python agent.py        # the agent - it decides which tools to call

llama3.2 calls tools reliably but its summaries can be generic. For sharper output, pull a stronger model and set OLLAMA_MODEL:

OLLAMA_MODEL=qwen3-coder:30b python agent.py

1. Ask the agent only about logs, then only about Docker — watch which tool it picks.
2. Add a third tool (e.g. return the last N lines of a log) and try it.
3. Change list_containers to also show stopped containers (docker ps -a).
4. Bonus: expose the agent behind a FastAPI /ask endpoint (reuse Day 1's pattern).