How to Deploy OpenMetadata with Docker: A Practical Tutorial

Deploying OpenMetadata with Docker

OpenMetadata ships official Docker images (openmetadata/server with over 4.9M pulls), making self-hosting straightforward. In this tutorial we walk through the core steps to get a working instance up and running — the same setup our tool page provides as a ready-made docker-compose template.

Step 1: Prerequisites

• Docker Engine 20.10+ and Docker Compose v2
• At least 2 vCPU and 4 GB RAM (4 vCPU / 8 GB recommended)
• Port 8080 free for the web UI

Step 2: Minimal docker-compose.yml

services:
  openmetadata:
    image: openmetadata/server:latest
    restart: unless-stopped
    ports:
      - 8080:8080
    volumes:
      - ./data/openmetadata:/data

Save the file, then run docker compose up -d. The server pulls the image, starts the application, and exposes the UI at http://localhost:8080.

Step 3: First-time setup

Open the UI, create the admin account, and connect your first data source — OpenMetadata includes native connectors for databases, warehouses, BI tools, and orchestration platforms like Airflow and dbt. Once connected, ingestion pipelines automatically populate the catalog with schemas, lineage, and profiles.

Performance expectations

Review: Is it worth it?

OpenMetadata stands out for combining catalog, lineage, quality, and governance in one open package with an active community. The Docker image is battle-tested (millions of pulls), the UI is polished, and the AI-context features are ahead of most competitors. The main trade-off is operational complexity at large scale — plan your backend services carefully.

Ready to try it? Head to our OpenMetadata tool page for the full docker-compose template, requirements, and installation resources.