Setting Up Postgres as a Data Source - GrowthBook Documentation

GrowthBook runs SELECT queries against your Postgres database to compute experiment results. Create a dedicated read-only user, allow network access from GrowthBook, then add the connection in Metrics and Data → Data Sources.

1. Create a read-only user

Use a dedicated role so GrowthBook cannot modify data. Grant SELECT only on the schemas that hold experiment assignment and metric events.

CREATE USER growthbook WITH PASSWORD 'use-a-strong-password';
GRANT CONNECT ON DATABASE your_database TO growthbook;
GRANT USAGE ON SCHEMA public TO growthbook;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO growthbook;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT SELECT ON TABLES TO growthbook;

Replace public with your analytics schema if events live elsewhere.

2. Allow network access

If you use GrowthBook Cloud, whitelist 52.70.79.40 in your Postgres firewall, security group, or pg_hba.conf. Self-hosted GrowthBook runs in Docker. To reach a database on the host machine, use host.docker.internal as the host instead of localhost.

3. Add Postgres in GrowthBook

Go to Metrics and Data → Data Sources.
Choose Postgres (guided setup if you use a supported event tracker, or manual setup for a custom schema).
Enter the connection fields:

Field	Notes
Host	Public hostname or IP. Cloud cannot reach `localhost`.
Port	Default `5432`.
Database	Used for the connection and the schema browser.
User / Password	The read-only user from step 1.
Default Schema	Optional. Prefixed onto generated SQL (for example `analytics.`).
Require SSL	Enable if your server requires TLS. You can attach a CA, client cert, and client key.

Save. GrowthBook runs a SELECT 1 query to confirm the connection.

4. Map assignments and metrics

After the connection succeeds, define experiment assignment queries and Fact Tables. Postgres supports the Schema Browser, so you can inspect tables and columns while writing SQL.