Environment:PacktPublishing LLM Engineers Handbook Docker MongoDB Qdrant Infrastructure

Knowledge Sources	LLM Engineers Handbook docker-compose.yml
Domains	Infrastructure, Databases
Last Updated	2026-02-08 08:00 GMT

Overview

Docker Compose environment running MongoDB (document store) and Qdrant (vector database) as local infrastructure services.

Description

This environment provides the two database backends required by the LLM Twin application. MongoDB serves as the NoSQL data warehouse for raw crawled documents (articles, posts, repositories) and user profiles. Qdrant serves as the vector database for cleaned documents, chunks, and embedded chunks used in the RAG pipeline. Both services run as Docker containers via `docker-compose.yml` with default credentials and port mappings.

Usage

Use this environment for all local development and testing. MongoDB is required for the Digital Data ETL and Feature Engineering workflows. Qdrant is required for Feature Engineering, Dataset Generation, and RAG Inference workflows. For cloud deployments, Qdrant Cloud can be used instead of the local container by setting `USE_QDRANT_CLOUD=True`.

System Requirements

Category	Requirement	Notes
Docker	>= 27.1.1	Docker Compose v2 included
Disk	~2GB free	For MongoDB and Qdrant data volumes
Network	Ports 27017, 6333, 6334 available	MongoDB and Qdrant default ports

Dependencies

Docker Images

`mongo:latest` (MongoDB)
`qdrant/qdrant:latest` (Qdrant vector database)

Python Packages (clients)

`pymongo` >= 4.6.2 (MongoDB driver)
`qdrant-client` >= 1.8.0 (Qdrant client)

Credentials

The following environment variables configure database access:

`DATABASE_HOST`: MongoDB connection string (default: `mongodb://llm_engineering:llm_engineering@127.0.0.1:27017`)
`DATABASE_NAME`: MongoDB database name (default: `twin`)
`QDRANT_DATABASE_HOST`: Qdrant host (default: `localhost`)
`QDRANT_DATABASE_PORT`: Qdrant port (default: `6333`)
`USE_QDRANT_CLOUD`: Boolean flag for Qdrant Cloud (default: `False`)
`QDRANT_CLOUD_URL`: Qdrant Cloud URL (required if `USE_QDRANT_CLOUD=True`)
`QDRANT_APIKEY`: Qdrant Cloud API key (required if `USE_QDRANT_CLOUD=True`)

Quick Install

# Start local infrastructure
docker compose up -d

# Verify services are running
docker compose ps

Code Evidence

MongoDB container configuration from `docker-compose.yml:2-17`:

mongo:
  image: mongo:latest
  environment:
    MONGO_INITDB_ROOT_USERNAME: "llm_engineering"
    MONGO_INITDB_ROOT_PASSWORD: "llm_engineering"
  ports:
    - 27017:27017

Qdrant container configuration from `docker-compose.yml:19-32`:

qdrant:
  image: qdrant/qdrant:latest
  ports:
    - 6333:6333
    - 6334:6334

MongoDB singleton connection from `llm_engineering/infrastructure/db/mongo.py:14`:

cls._instance = MongoClient(settings.DATABASE_HOST)

Qdrant connection with cloud fallback from `llm_engineering/infrastructure/db/qdrant.py:14-27`:

if settings.USE_QDRANT_CLOUD:
    cls._instance = QdrantClient(
        url=settings.QDRANT_CLOUD_URL,
        api_key=settings.QDRANT_APIKEY,
    )
else:
    cls._instance = QdrantClient(
        host=settings.QDRANT_DATABASE_HOST,
        port=settings.QDRANT_DATABASE_PORT,
    )

Common Errors

Error Message	Cause	Solution
`Connection refused` on port 27017	MongoDB container not running	Run `docker compose up -d`
`Connection refused` on port 6333	Qdrant container not running	Run `docker compose up -d`
`Authentication failed` for MongoDB	Wrong credentials	Verify `DATABASE_HOST` matches docker-compose credentials

Compatibility Notes

Qdrant Cloud: Set `USE_QDRANT_CLOUD=True` with `QDRANT_CLOUD_URL` and `QDRANT_APIKEY` to use managed Qdrant instead of local container.
Docker Desktop: On macOS/Windows, Docker Desktop must be running before `docker compose up`.
Port Conflicts: If ports 27017 or 6333 are already in use, modify `docker-compose.yml` and update the corresponding settings.

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment