Intermingling your pre-existing roles in Snowflake with Immuta can be confusing at first. This guide outlines some best practices on how to think about roles in each platform.
Roles play a crucial role in Snowflake by organizing and controlling access to data, platform permissions, and data warehouses. Immuta also leverages Snowflake roles to grant users permission to read data based on subscription policies.
Users who consume data (directly in Snowflake or through other applications) need roles to access objects. But roles are also used to control write, Snowflake warehouses, and Snowflake permissions through system-defined roles.
To manage this at scale, Immuta recommends taking a 4-layer approach, where you separate the different permissions into different roles:
Roles for read access (Immuta managed)
Roles for write access (optional, soon supported by Immuta)
Roles for warehouse, internal billing
Roles for Snowflake permissions (optional)
Read access is managed by Immuta. By using subscription policies, data access can be controlled to the table level. Attribute-based table GRANTS help you scale compared to RBAC, where access control is typically done on a schema or database level.
Since Immuta leverages Snowflake roles, you can still use existing roles in Snowflake. This means you can gradually migrate to an Immuta-protected Snowflake.
Write access is typically granted on a schema or database. This makes it easy to manage in Snowflake through manual grants. We recommend creating roles that give insert, update, and delete permissions to a specific schema or database and attach this role to a user. This attachment can be done manually or using your identity manager groups. (See the Snowflake documentation for details.) Note that Immuta is working towards supporting write policies, so this will not need to be separately managed for long.
Warehouses are granted to users to give them access to computing resources. Since this is directly tied to Snowflake’s consumption model, warehouses are typically linked to cost centers for (internal) billing purposes. Immuta recommends creating a role/warehouse per team/domain/cost center and granting this warehouse role to users using identity manager groups.
Snowflake permissions are granted through system-defined roles like ACCOUNTADMIN or SECURITYADMIN. These are high-privilege roles that are only granted to administrators. This can be done manually or using AD groups.
Snowflake allows users to select a specific role, but you can also use all roles simultaneously. Immuta recommends using all roles since that helps to separate the different roles.
This feature is called ‘secondary roles’ and can be enabled using the following command in Snowflake: USE SECONDARY ROLES ALL
Alternatively, you could create personal roles and grant the warehouse-role/immuta-read-role and possibly the snowflake-permission-role and write-role to this.
Immuta has two types of service accounts to connect to Snowflake:
Policy role: This role gives Immuta the power to create and apply policy. Immuta can create this policy role automatically, or you can run the provided bootstrap script manually to create the policy role.
Data ownership role: This role is used to register data sources. A service account/role is recommended so that when the user moves or leaves the organization, Immuta will still have the proper credentials to connect to Snowflake. You can follow one of the two best practices:
A central role for registration (recommended): It is recommended that you create a service role/user with USAGE permissions for all objects in Snowflake. This allows Immuta to register all the objects from Snowflake, populate the Immuta catalog, and scan the objects for sensitive data using Immuta Discover. Immuta will not apply policy directly by default, so no existing access will be impacted.
A role per team/domain (alternative): Alternatively, if you cannot create a role with USAGE permissions for all objects, you can allow the different domains or teams in the organization to use a service user/role scoped to their data to register data sources. This is delegating metadata registration and aligns well with data mesh type use cases and means every team is responsible for registering their data sets in Immuta.