Simple RAG pipeline with OpenLLMetry tracing

A sample RAG pipeline built with LlamaIndex and ChromaDB as a vector store auto instrumented with OpenLLMetry

Technologies used:

• Python
• uv
• Docker
• ChromaDB
• LLamaIndex
• OpenAI API
• OpenLLMetry
• Datadog

How to Run

Run ChromaDB and Datadog agent with Docker Compose:

1. Set Datadog API KEY

export DD_API_KEY=<my-datadog-api-key>

2. Run Docker Compose services (daemon mode)

docker compose up -d

Run the pipeline

1. Set OpenAI API key

export OPENAI_API_KEY=<my-openai-api-key>

2. Set OTel backend via TRACELOOP_BASE_URL env var (that'll be picked up by Traceloop SDK)

export TRACELOOP_BASE_URL=http://localhost:4318

3. Run the pipeline with uv

uv run main.py