Connecting Looker (Google Cloud core) to your database

Once your Looker (Google Cloud core) instance has been provisioned, it is listed on the Instances page of your Google Cloud project. Click the instance URL to access and authenticate in to the instance.

Once you have logged in to your Looker (Google Cloud core) instance, you can set up a database connection to your Looker (Google Cloud core) instance.

Set up a database connection

Looker (Google Cloud core) must be connected to a database to enable data exploration. See the list of supported dialects to learn which dialects are supported by Looker (Google Cloud core).

You can create a database connection within a Looker (Google Cloud core) instance if you have one of the following permissions:

You can follow the Set up Looker guide that appears dynamically within the Looker (Google Cloud core) instance to connect your database, or follow the steps listed on the Connecting Looker to your database documentation page and the dialect-specific documentation pages.

If your Looker (Google Cloud core) instance uses a private IP connection, you must set up a route or a private connection to connect it to any of the following types of databases:

  • A database in a different network within Google Cloud
  • A database that is hosted by another cloud service provider
  • An on-premises database

Learn more about private networks and external services on the Private IP networking with Looker (Google Cloud core) documentation page.

Once a database connection is set up, you are ready to set up a LookML project.

Using Application Default Credentials to connect to a BigQuery database

Looker (Google Cloud core) instances can use Application Default Credentials (ADC) to authenticate when you're setting up a connection to a BigQuery Standard SQL database. When you use ADC, the connection will authenticate to the database by using the credentials of the Looker (Google Cloud core) project's service account.

To use ADC with a BigQuery database, select Application Default Credentials in the Authentication field of the Connection Settings page of the Looker instance. For the full procedure, see the documentation for connecting Looker to a BigQuery database.

If you're connecting to a BigQuery database that is in a different project from your Looker (Google Cloud core) instance, some additional setup is required. See the Using Application Default Credentials with a BigQuery database in a different Google Cloud project section.

Service account impersonation

If you want to authenticate to the BigQuery database by using a service account other than the Looker (Google Cloud core) project's service account, you can create a delegated request flow by entering another service account, or a comma-separated chain of service accounts, into the Impersonated Service Account field. The Looker (Google Cloud core) service account is automatically used as the first service account in the chain and does not need to be added to the field. The last service account in the chain (also known as the impersonated service account) is the one that authenticates with the database.

When using service account impersonation, do the following:

Using Application Default Credentials with a BigQuery database in a different Google Cloud project

The steps for using ADC for a BigQuery Standard SQL database that is outside the project that houses your Looker (Google Cloud core) instance are the same as those for setting up a connection inside the same project. However, prior to setting up the connection in your Looker (Google Cloud core) instance, your Looker (Google Cloud core) project's service account must have the following IAM roles:

If the Looker (Google Cloud core) service account doesn't already have IAM roles in the project that contains the BigQuery dataset, use the service account's email address when granting roles in that project. To find the service account's email address, go to the IAM page in the Google Cloud console and select the Include Google-provided role grants checkbox. The email will have the format service-<project number>@gcp-sa-looker.iam.gserviceaccount.com. Use that email to grant the proper roles to the service account.

Once the proper roles are granted, follow the steps to use ADC.

You can now use ADC with this BigQuery Standard SQL database. The project attached to the service account that is specified in the Connection Settings page will be used for billing and also act as the default project.

Using Application Default Credentials to connect to a Cloud SQL database

Looker (Google Cloud core) instances can use ADC to authenticate a connection to a Cloud SQL database (either Cloud SQL for PostgreSQL or Cloud SQL for MySQL). When you use ADC to authenticate into your Cloud SQL database, the Google Cloud project where the Cloud SQL database is running is the project that is billed for Looker queries.

For Looker connections to Cloud SQL that use ADC, ADC impersonates a service account or a chain of service accounts to access your database. When you create the Looker connection to your database, you use the IAM database username(s) field to specify the service account, or the chain of service accounts, that ADC will impersonate. The Looker service account that was created automatically when you created the Looker (Google Cloud core) instance is automatically used as the first service account in the chain and does not need to be added to the field.

If you want to authenticate to your Cloud SQL database by using a service account other than the Looker service account, you can create a delegated request flow by entering another service account, or a comma-separated chain of service accounts, into the IAM database username(s) field.

The last service account in the chain (also known as the impersonated service account) is the one that authenticates with the database, and this account must be added as a user on your Cloud SQL database. If you are using the Looker service account as the last service account in the chain (by leaving the IAM database username(s) field blank), you must add the Looker service account as a user on your Cloud SQL database.

The following are the general steps for connecting a Cloud SQL for PostgreSQL or Cloud SQL for MySQL database to Looker using ADC:

  1. Add the impersonated service account to your Cloud SQL database.
  2. Set up service account impersonation on your Cloud SQL database.
  3. Connect to your database to run additional configuration commands for Cloud SQL for PostgreSQL or Cloud SQL for MySQL.
  4. Create the Looker connection to your database.

Add the impersonated service account to your Cloud SQL database

When you create the Looker connection to your database, you use the IAM database username(s) field to specify the service account, or the chain of service accounts, that ADC will impersonate to perform actions on your database. The last service account in the impersonation chain is considered the impersonated service account.

To use ADC with Cloud SQL, you must add the impersonated service account to your Cloud SQL database:

  • In the default case, if you leave the IAM database username(s) field blank, ADC will impersonate the Looker service account. In this case, the Looker service account is the impersonated service account, so you need to add the Looker service account to your Cloud SQL database. See the Create a Looker (Google Cloud core) instance documentation page for information about the Looker service account and for the procedure for viewing the Looker service account email address.
  • If you specify a service account other than the Looker service account, or if you specify a chain of service accounts in the IAM database username(s) field, you must add the last service account in the impersonation chain to your Cloud SQL database.

To add a service account to your Cloud SQL database, you must have the Cloud SQL Admin IAM role.

Follow the "Add an IAM user or service account to your database instance" procedure for your database dialect to add the impersonated service account to your Cloud SQL database:

Set up service account impersonation on your Cloud SQL database

Once you have created the Cloud SQL user on your database, you must set up your Cloud SQL database for service impersonation by performing the following steps:

  1. Follow the procedure to enable the Cloud SQL Admin API.
  2. Make sure that all service accounts in the chain, including the Looker service account, have the appropriate IAM permissions.

  3. Follow the procedure for granting a single role in the Google Cloud console. Grant the following Cloud SQL roles to the impersonated service account that you added to your Cloud SQL database:

    If you specify a service account other than the Looker service account, or if you specify a chain of service accounts in the IAM database username(s) field, grant every service account in the chain the following permission:

Additional configuration commands for Cloud SQL for MySQL

For Cloud SQL for MySQL, you must connect to your database instance and run the following command on the Cloud SQL for MySQL database:

GRANT ALL on DATABASE_NAME.* to 'DATABASE_USER'@'%'

Replace the following:

  • DATABASE_NAME: The name of your database.
  • DATABASE_USER: The truncated service account username for the impersonated service account that you added to your Cloud SQL database. The service account will have the format service-<project number>@gcp-sa-looker.iam.gserviceaccount.com. Truncate the username by removing the @ and everything that follows. After truncating, the username would look like service-<project number>.

For example, if the service account username is service-12345678901@gcp-sa-looker.iam.gserviceaccount.com and the database name is looker-test, the command would be as follows:

GRANT ALL on looker-test.* to 'service-12345678901'@'%'

Additional configuration commands for Cloud SQL for PostgreSQL

For Cloud SQL for PostgreSQL, you must connect to your database instance and run some configuration commands on the Cloud SQL for PostgreSQL database:

  • Grant the user permissions on your database as described in the Users and security section of the PostgreSQL documentation page.
  • Set the search path for the Looker SQL Runner to use to retrieve metadata from your database, as described in the Setting the search_path section of the Looker documentation PostgreSQL page.

Create the connection to Looker (Google Cloud core)

To create the connection from Looker to your database, follow these steps:

  1. In the Admin section of Looker, select Connections, and then click Add Connection.
  2. From the Dialect drop-down menu, select Google Cloud PostgreSQL or, for Cloud SQL for MySQL, select Google Cloud SQL.
  3. In the Authentication section, click the Application Default Credentials option.

  4. In the IAM database username(s) field, specify the service account, or the chain of service accounts, that you want ADC to impersonate to perform actions on your database:

  5. Fill out the rest of the connection details. The majority of the settings are common to most database dialects. See the Connecting Looker to your database documentation page for information.

  6. To verify that the connection is successful, click Test. See the Testing database connectivity documentation page for troubleshooting information.

  7. To save these settings, click Connect.

Supported dialects for Looker (Google Cloud core)

The following table shows the Looker (Google Cloud core) support for database dialects:

Dialect Supported?
Actian Avalanche
No
Amazon Athena
Yes
Amazon Aurora MySQL
Yes
Amazon Redshift
Yes
Apache Druid
No
Apache Druid 0.13+
No
Apache Druid 0.18+
Yes
Apache Hive 2.3+
No
Apache Hive 3.1.2+
Yes
Apache Spark 3+
Yes
ClickHouse
Yes
Cloudera Impala 3.1+
Yes
Cloudera Impala 3.1+ with Native Driver
No
Cloudera Impala with Native Driver
No
DataVirtuality
No
Databricks
Yes
Denodo 7
No
Denodo 8
Yes
Dremio
No
Dremio 11+
Yes
Exasol
No
Firebolt
No
Google BigQuery Legacy SQL
No
Google BigQuery Standard SQL
Yes
Google Cloud PostgreSQL
Yes
Google Cloud SQL
Yes
Google Spanner
Yes
Greenplum
No
HyperSQL
No
IBM Netezza
Yes
MariaDB
Yes
Microsoft Azure PostgreSQL
Yes
Microsoft Azure SQL Database
Yes
Microsoft Azure Synapse Analytics
Yes
Microsoft SQL Server 2008+
No
Microsoft SQL Server 2012+
No
Microsoft SQL Server 2016
No
Microsoft SQL Server 2017+
Yes
MongoBI
No
MySQL
No
MySQL 8.0.12+
Yes
Oracle
Yes
Oracle ADWC
No
PostgreSQL 9.5+
Yes
PostgreSQL pre-9.5
No
PrestoDB
Yes
PrestoSQL
Yes
SAP HANA 2+
Yes
SingleStore
No
SingleStore 7+
Yes
Snowflake
Yes
Teradata
No
Trino
Yes
Vector
No
Vertica
Yes

Database configuration instructions

Instructions are available for these SQL dialects:

What's next