Skip to content
For the complete documentation index, see llms.txt. This page is also available as Markdown.

Domains

Learn about how domains allow you to delegate access management and how Immuta enforces domain policies

Domains are containers of data sources that let you assign data ownership and access management to specific business units, subject matter experts, or teams. Instead of routing every governance decision through a central team, domains let the people closest to the data make access decisions for it, while central governance keeps oversight across the organization.

Within a domain, users are assigned domain permissions to complete the following actions:

  • Manage policies

  • Audit activity

  • Manage identifiers

  • Publish data products

Because these permissions are scoped to specific domains, users only have power over the data in the domains you assign them, not all data in your organization.

A data owner registers data in Immuta as data sources. A user with the GOVERNANCE permission creates a domain and assigns data sources to it, either manually or dynamically based on tags. A user with the USER_ADMIN permission then grants domain permissions to the users or groups who will steward that domain. From then on, those stewards can author policies, audit activity, manage identifiers, and publish data products for the data sources in their domain only. Global policies set by central governance continue to apply to every data source, regardless of domain.

Domain data sources

Data sources can be assigned to domains based on business units in your organization or any other method that suits your business goals and policy management strategy. Data sources can be assigned to a domain one of two ways:

Assignment method
How it works
Best for

Manual

The GOVERNANCE user selects each data source and adds or removes it through the UI or API. Nothing changes without user intervention.

Small, stable sets of data sources that rarely change.

Dynamic

The GOVERNANCE user selects one or more table tags. Data sources with any of those tags are added to or removed from the domain automatically. Column tags cannot be used.

Large or growing data estates. Dynamic assignment scales better and needs less maintenance.

For the most hands-off setup, use tags that are applied automatically, such as connection tags or tags from an external catalog. New data sources then land in the correct domain as soon as they are onboarded.

Once a data source is assigned to a domain (or multiple domains), users with a domain permission can start performing actions on those data sources.

See the Create and manage domains guide for instructions on creating a domain and adding data sources to it.

Dynamic domain assignment

See how data sources are assigned dynamically below.

Dynamic assignment process

Auto-assignment is implemented with many Immuta features and components:

  1. A user with the GOVERNANCE permission creates a domain.

  2. During creation, the user decides if data sources should be added dynamically or manually. The user picks dynamically in this example.

  3. The tag(s) for dynamic assignment are selected by the user.

  4. The domain is created; it has all the data sources that already exist with the selected tag(s).

  5. New data sources are created through schema monitoring or object sync:

    1. Data sources created from object sync automatically get a tag from their connection.

    2. Immuta connects them to the external catalog, pulling in the appropriate tags.

  6. Once the data sources get any of the domain assignment tags, they are automatically added to the domain with the selected, matching tag.

  7. Tags may change on existing data sources:

    1. Updated by a user: If a user adds or removes the domain assignment tag from a data source, the data source will be added or removed from the domain automatically (within a few seconds).

    2. Updated in an external catalog: If changes come from an external catalog, those changes will be pulled into Immuta every 24 hours. Any changes to tags, and therefore domain assignment, will be enacted within that timeframe.

Performance

Adding data sources to a domain, whether dynamically with tag assignment or manually using bulk actions, will take longer the more data sources you are adding. The following scenarios have been tested in order to provide a benchmark of the average time to complete the action.

Scenario 1

Add 1,000 data sources to a domain.

<1 second on average

Scenario 2

Add 10,000 data sources to a domain.

2 seconds on average

Scenario 3

Add 100,000 data sources to a domain.

30 seconds on average

Domain policies

Whether created within or outside a domain, Immuta policies enforce access as usual: users who meet the restrictions outlined in the policy applied to a data source may access that data source. However, domains restrict who can author policies that apply to data sources assigned to that domain. This policy management restriction gives organizations more control over how much power governance users have on data. Furthermore, it can make Immuta easier to use by allowing more people to author policies.

Users with the Manage Policies permission in a domain can set global policies to apply to those domains. In the example below, the domain policy manager can only apply global policies to the HR domain:

If that same user had the Manage Policies permission for the HR and Marketing domains, they could write global policies to apply to data sources within both of those domains. If that user had the Manage Policies permission on all domains in their organization or the global GOVERNANCE permission, they could set global policies to apply to all data sources.

When a data source is removed from or added to a domain, Immuta recomputes the policies that apply to the domain. Any policies associated with a domain will be added to or removed from a data source when it is added to or removed from the domain.

See the Create and manage domains guide for instructions on authoring domain policies.

Designing domain policies

When integrating domains to distribute policy management across your organization, data governance and access control must be applied horizontally (globally across data in your organization) and vertically (locally within specific domains or data products).

Global policies should reflect your organization's broadest principles, regardless of domain. Domain policies should be fine-grained and specific to the context or use case of that domain.

Policy level
Authored by
Applies to
Example

Global

Users with the GOVERNANCE permission

All data sources in the organization

Mask all PII across the organization.

Domain

Users with the Manage Policies permission on the domain

Data sources in the domains where the user has that permission

Only show rows in the Sales table where the country column matches the user's office location.

Local

Data owners

A single data source

Disable a domain policy that conflicts with a more relevant local policy.

Since policies apply to domain data sources at multiple levels, there are instances in which two or three policies could apply to a single domain data source. Data owners can disable and enable policies that are most appropriate for their data source when a conflict occurs. For details about policy conflicts, merges, and policy conflict management, see the following pages:

Identifiers in domains

Identifiers are grouped into domains with data sources, and when identification runs, a domain's identifiers are only checked against the data sources in that domain. Users with the Manage Identifiers permission can create domain-specific identifiers or add them from the reference identifiers. This lets stewards tailor identification to the unique data types, tagging standards, and risk levels in their domain (such as Finance, HR, or Clinical), and manage identification logic independently, without impacting global or unrelated domain logic.

Identification runs automatically when a data source is added to a domain with identifiers. Combined with dynamic data source assignment, this means new data sources are tagged as soon as they are onboarded. See the Data identification reference guide for details.

Domains in the Request app

Domains determine how data products are organized and published in the Request app. Users with the Manage Data Products permission on a domain can publish data products from that domain's data sources, and domain permissions can be used as a steward source in review flows. A well-structured domain model makes it easier for users to find and request the data they need.

Use cases

Delegate policy management by line of business

A large financial services firm has two business units with very different data: insurance data managed by actuaries and investment data owned by portfolio managers. Central governance wants each unit to manage policies for its own data without gaining access to the other unit's data.

This can be configured with domains:

  1. Create an Insurance domain and an Investment domain as a user with the GOVERNANCE permission.

  2. Assign data sources dynamically using a business unit tag from your external catalog, such as Business Unit.Insurance and Business Unit.Investment.

  3. Grant the Manage Policies domain permission to the actuaries group on the Insurance domain and to the portfolio managers group on the Investment domain.

  4. Author domain policies within each domain for that unit's specific access needs.

Each business unit governs its own data, while global policies, such as masking PII, continue to apply across both domains.

Organize domains by data platform with connection tags

A tech-focused organization has separate teams for Snowflake and Databricks: Snowflake is stewarded by platform admins and Databricks by a data science governance team. New tables are registered constantly through object sync.

This can be configured with domains:

  1. Create a Snowflake domain and a Databricks domain as a user with the GOVERNANCE permission.

  2. Assign data sources dynamically using each connection's connection tag.

  3. Grant the Manage Policies domain permission to the platform admins on the Snowflake domain and to the data science governance team on the Databricks domain.

New data sources created through object sync automatically receive their connection tag and land in the correct domain, with no manual assignment.

Meet regional compliance requirements

A global enterprise must apply EU-specific compliance rules to data from its EMEA operations, and the EU data compliance team needs to show regulators who accessed that data.

This can be configured with domains:

  1. Create an EMEA domain as a user with the GOVERNANCE permission.

  2. Assign data sources dynamically using a region tag, such as Region.EMEA.

  3. Grant the Manage Policies and Audit Activity domain permissions to the EU data compliance team.

  4. Author domain policies for EU-specific requirements, such as masking additional fields considered personal data under GDPR.

  5. Use the domain's Audit tab to review queries and policy changes when preparing for regulatory audits.

The compliance team enforces and audits regional rules without changing policies that apply to the rest of the organization.

Tailor identification for a clinical domain

A healthcare organization wants new clinical tables tagged as soon as they are registered. The built-in identifiers catch common data like Social Security numbers, but they don't recognize the organization's medical record number (MRN) format.

This can be configured with domains:

  1. Create a Clinical domain and assign data sources dynamically using the clinical database's connection tag.

  2. Grant the Manage Identifiers domain permission to the clinical data stewards.

  3. Add relevant reference identifiers to the domain, such as the built-in Social Security number identifier.

  4. Create a domain-specific MRN identifier with a regex criteria that matches the MRN format and applies an MRN tag.

Audit

Users with the global AUDIT permission can see all audit events throughout Immuta. To scope this audit information to just a single domain, users can be given the domain Audit Activity permission.

Domain events

The following domain-related events are audited and can be found on the audit page in the UI:

Data source assignment

The following events related to data source assignment are audited and can be found on the audit page in the UI:

  • DomainDataSourcesUpdated: Data sources are assigned to or removed from the domain.

  • TagApplied: A tag is applied to a data source or column, which could lead to a data source being added to a domain.

  • TagRemoved: A tag is removed from a data source or column, which could lead to a data source being removed from a domain.

Audit tab

If a user has the domain Audit Activity permission on a particular domain, they can see domain-specific audit events in the Audit tab of the domain. The following domain activities are in that tab:

  • Domain: Events from when the domain was created, permissions were updated, and the domain was updated.

  • Data sources: Events from the data sources within the specific domain.

  • Policies: Events for policies created, edited, or deleted that affect data sources within the domain.

  • Queries: Events for user queries against data sources within the domain.

Permissions

Domain permissions can be added to users or groups by a user with the USER_ADMIN Immuta permission. See the Immuta users and permissions guide for a list of global and domain permissions necessary to manage domains.

Last updated

Was this helpful?