Deploy and Host self hosted Metabase Databricks (Open-Source BI for Databricks) on Railway

This template runs self-hosted Metabase on Railway with a Postgres 18 service holding Metabase's app database. Databricks isn't wired in at deploy time; you add it later through the Admin UI.

About Hosting Metabase Databricks open-source software on Railway (self hosted Metabase template)

Your dashboards read from the Databricks lakehouse, but the Metabase that renders them keeps its own state in Postgres. Databricks is added later in the admin UI, not as a template variable.

Railway runs the server, volume and Postgres lifecycle for you.

Why Deploy Metabase Databricks, the Databricks AI/BI Dashboards alternative on Railway (Railway Free Trial)

Databricks AI/BI Dashboards suits teams already working inside the workspace, but it ties your dashboards to one platform. You own upgrades, backups, and uptime.

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Metabase Databricks on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.

Railway vs Other Hosting Providers and VPS for Metabase Databricks self hosting

Railway gives you Metabase and its Postgres 18 app database in one project, while a VPS leaves OS patching, TLS, and backups on your plate.

Provider	What You Get with Railway	What You Get with the Other Provider
DigitalOcean	Metabase and Postgres 18 from one template, with a public domain and volumes managed in one dashboard.	Full root access and fixed droplet pricing, but you install Docker, set up a reverse proxy, and back up Postgres yourself.
AWS	A Metabase Databricks instance serving HTTPS without VPC, IAM, or security group setup.	Deep networking control and managed RDS options, with more line items to watch on the invoice.
Hetzner	Redeploys from the dashboard, with no SSH or firewall rules to manage.	Low raw compute cost per GB of RAM, which can win at scale if you like running the server yourself.

Common Use Cases for hosted Metabase Databricks

Executive dashboards on lakehouse tables. Finance wants weekly revenue and churn from the gold tables data engineering already maintains. Connect one SQL warehouse, scope the catalog with Only these..., and publish dashboards that refresh straight from Databricks.
Ad hoc questions for non-SQL users. Product managers can summarize and filter visually in the question builder, while analysts keep writing native SQL against the same warehouse.
Scoped access by schema. When one workspace holds marketing, billing, and experiments, use All except... to keep sensitive schemas out of the sync. Metabase only sees what you allow.
Slow warehouse triage. Turn Re-run queries for simple explorations off so people click Run before summarizations apply, and keep Periodically refingerprint tables off to avoid extra scans on every sync.

Dependencies for Metabase Databricks Docker hosted on Railway

Metabase Databricks on Railway is two services: a Metabase container and a Postgres database holding Metabase's own application data.

Deployment Dependencies for Managed Metabase Databricks Service (BI for Databricks SQL Warehouses)

The template runs the metabase/metabase image (tag latest, port 3000) alongside a Railway Postgres 18 service with a volume.

Implementation Details for Metabase Databricks (Using Metabase official docker image)

After first-boot setup, open the grid icon, then Admin, then Databases, then Add a database, and choose Databricks. Save after every edit.

How does Metabase Databricks compare against other BI tools for Databricks platforms

Picking a BI tool for a Databricks lakehouse usually comes down to where your analysts already work and what you want to pay for seats.

Metabase Databricks vs Databricks AI/BI Dashboards (Databricks AI/BI Dashboards Alternative)

Deployment: Metabase runs in your own Railway project and connects to the warehouse over the Databricks driver. AI/BI Dashboards lives inside the workspace with nothing extra to host.
Cost model: The OSS edition is free apart from Railway compute and Postgres. Databricks bills SQL warehouse time either way.
Where Databricks wins: Native Unity Catalog governance and zero extra infrastructure.

Metabase Databricks vs Tableau (Tableau Alternative)

Strength: Tableau's visual depth and polish remain best in class for highly designed analytical storytelling.
Cost model: Tableau licenses per creator and per viewer, which climbs fast in a large org. Metabase OSS has no seat fee.
Where Tableau wins: Advanced visual customization and mature enterprise governance the open-source edition does not match.

Metabase Databricks vs Power BI (Power BI Alternative)

Strength: Power BI fits Microsoft-heavy shops with tight Excel and Teams integration.
Cost model: Power BI Pro is per user, while self-hosted Metabase swaps that for Railway bills that stay small for a modest team.
Where Power BI wins: Deep Microsoft ecosystem ties and a familiar interface for Excel-first users.

Metabase Databricks vs Apache Superset (Apache Superset Alternative)

Similarity: Superset is also self-hostable and open source, so the infrastructure story is close.
Difference: Metabase offers a friendlier question builder for non-SQL users, while Superset gives more control over custom charts and SQL-first workflows.
Where Superset wins: Richer chart types for data engineers who prefer writing queries directly.

How to use Metabase Databricks (the OSS BI for Databricks SQL Warehouses)?

Once the deploy is green and the setup wizard is done, point Metabase at your lakehouse. Click the grid icon top right, then Admin Databases Add a database and pick Databricks.

Under Authentication, a personal access token works for a quick test, but a service principal over OAuth M2M is the better long-term choice. Save, then let the first sync finish.

How to self host Metabase Databricks on other VPS Services (Metabase Databricks self hosting guide)

A VPS run is the Railway setup minus the plumbing Railway does for you. You still need Postgres for Metabase's own state, and the Databricks connection has to be right or the form stays stuck.

Clone the Repository

git clone https://github.com/metabase/metabase.git your-domain-metabase
cd your-domain-metabase

Install Dependencies

You need Docker and a Postgres instance the container can reach. Keep the app database separate from Databricks, since the warehouse is only a data source.

docker run -d --name metabase-db -e POSTGRES_PASSWORD=change-me postgres:18

Configure Environment Variables

Set the app database variables and the public site URL, plus a stable encryption key before saving any Databricks credentials. Never rotate that key.

MB_DB_TYPE=postgres
MB_DB_HOST=metabase-db
MB_DB_PORT=5432
MB_DB_USER=postgres
MB_DB_PASS=change-me
MB_DB_DBNAME=metabase
MB_SITE_URL=https://your-domain
MB_ENCRYPTION_SECRET_KEY=your-generated-key

Start the Metabase Databricks Application

Start the container and finish the setup wizard. Then open the grid icon, go to Admin, Databases, Add a database, and pick Databricks. Railway handles TLS and restarts for you, which a VPS does not.

Official Pricing of Metabase Databricks (Metabase Databricks pricing)

The software is free.

Metabase Cloud, if you'd rather not run it yourself, is priced as follows. Pro is $575 per month with 10 users included and $12 per extra user. Enterprise starts at $20,000 per year.

Metabase Databricks cloud vs self hosted comparison (Pricing, features, costs, and more)

Self hosting splits the bill in two. Metabase and its Postgres sit on your Railway invoice, while queries run on a Databricks SQL warehouse you already pay for. The OSS build has no license fee.

Monthly cost of self hosting Metabase Databricks on Railway

A small Metabase plus Postgres stack on Railway usually runs about $5 to $15/month on the Hobby plan.

System Requirements for Hosting Metabase Databricks on a VPS

Plan for at least 2 GB of RAM and a persistent volume for Metabase's Postgres application database, since H2 is wiped on redeploy.

Frequently Asked Questions (FAQs)

Which Databricks host value do I paste into Metabase?

Use only the Server Hostname from your compute resource, such as xxxxxxxxxx.cloud.databricks.com. Leave off any https:// prefix or path.

Personal access token or service principal?

A personal access token is quick for a small team, but it ties access to one person. A service principal with OAuth M2M suits shared dashboards better, since it outlives any single employee's account.

Why does Metabase ask for a default catalog?

Native SQL questions run without catalog qualification once a default is set, so analysts skip typing full three-part names.

How do I limit which catalogs and schemas sync?

Turn on Enable multiple catalogs, then use Catalogs and schemas with Only these... or All except.... Only the * wildcard works in those lists.

Will dashboards keep my SQL warehouse running?

Possibly. Frequent syncs and refingerprinting send queries to the warehouse, so tune schedules and check auto-stop behavior in Databricks documentation.

Where should I store the encryption key?

Set MB_ENCRYPTION_SECRET_KEY on the Metabase service before saving any warehouse credentials, and never change it afterward. Changing it leaves stored connection secrets unreadable.