home User Guide Getting Started Help Center Documentation Community Training Certification
Looker keyboard_arrow_down
language keyboard_arrow_down
PostgreSQL on Heroku

Getting Credentials from Heroku

To connect to your PostgreSQL database on Heroku, you’ll need the following information, which you’ll input in the Connections page of Looker:

Encrypting Network Traffic

Looker strongly recommends encrypting network traffic between the Looker application and your database. In the case of Heroku Postgres, the only option available is to use SSL encryption. SSL must be enabled in your Looker connection in order for Heroku Postgres to function properly.

Users and Security

Looker will use the default user and password provided by Heroku to access the database. New users cannot be added to the Heroku Postgres instance, per their documentation.

Choosing the Type of Analytics Database

We never recommend connecting Looker directly to a production database. Instead, choose one of the two following options:

Streaming Replica “Hot Swap” Connection

Forked Version of the Production DB
(requires development resources from your organization)

If you are using Postgres version earlier than 9.2, you will not be able to kill queries through your database’s SQL console or through Looker

Temp Schema Setup (forked databases only)

Create a schema owned by the default user on the forked database:

CREATE SCHEMA looker_scratch AUTHORIZATION <default_user>;

This scratch schema will be used for persistent derived tables. Please note this will only function properly when you’re using a forked version of your database. You must NOT be using Looker on top of a “hot swap” streaming replication database, because you must be able to write to the database from outside the application.

Feature Support

For Looker to support some features, your database dialect must also support them.

In Looker 21.0, PostgreSQL supports the following Looker features:

Next Steps

After completing the database configuration, you can connect to the database from Looker using these directions.