Hybrid Manager (HM) has two types of users: human users (native users or IDP users) and machine users.
You can assign any of the default roles to both types of users, except that only machine users can be estate ingesters.
The two main categories of roles are organization-level and project-level.
Organization-level roles
You can assign these roles to human or machine-users using the User Management option of your user profile menu at the top-right of the navigation bar in the HM console.
Organization Administrator
This role provides read access at the organization level, allowing the user to view information and settings in the organization as a whole.
Organization Owner
The purpose of this role is to manage the organization at a high level. Organization owners can create projects and assign organization-level roles to other users.
Platform Administrator
This role is for users who need to access and manage the underlying platform components of HM. They can access platform management and monitoring tools.
AI Model Manager
This role is for users who work with the AI model and model service feature of HM, granting them full access to its functions. It also grants access to the Agent Studio flow builder UI.
Appliance Administrator
This role applies only to on-premises HM deployments. It grants access to appliance-level configuration and monitoring — that is, the underlying appliance hardware and software that hosts HM, as distinct from the HM platform itself. For details, see your on-premises deployment documentation.
Project-level roles
These roles can be assigned to human or machine-users (except for Estate Ingester). Select Users in the left navigation when viewing a project.
Project Owner
In a specific project, the Project Owner role has the highest level of authority. They can take all actions in that project and are responsible for assigning project-level roles to other users.
Project Editor
This role is for users who need to actively work with the data in a project. Users with the Project Editor role have read and write access to data.
Project Viewer
For users who only need to observe the data in a project, the Project Viewer role provides read-only access.
Estate Ingester
This role is for machine-users and provides access to perform estate ingests in a project.
Pipeline Editor
This role is for users who need to create and manage AI pipelines and knowledge bases in a project.
Catalog Data Reader
Users with this role can read the Iceberg namespace/table/view of project-scoped catalogs.
Catalog Data Writer
This role allows users to read, write, and delete the Iceberg namespace/table/view of all project-scoped catalogs in a project.
Accessing a catalog's data access using the Iceberg REST API
Both first-party and third-party Iceberg REST clients must authenticate using an HM access key. Set one of these HTTP headers:
x-access-key: <access-key-value> Authorization: Bearer <access-key-value>
Note
Bearer is case-sensitive.
Warning
HM explicitly denies Iceberg endpoint access via UI session cookies. Access keys are required.
The following table maps each Iceberg REST API operation to the HM permission and minimum role it requires.
| Iceberg API | Required HM permission | Minimum role |
|---|---|---|
| list_namespace | read:catalog-data | Catalog Data Reader |
| create_namespace | read:catalog-data, write:catalog-data | Catalog Data Writer |
| load_namespace_metadata | read:catalog-data | Catalog Data Reader |
| namespace_exists | read:catalog-data | Catalog Data Reader |
| drop_namespace | write:catalog-data | Catalog Data Writer |
| update_namespace_properties | write:catalog-data | Catalog Data Writer |
| list_tables | read:catalog-data | Catalog Data Reader |
| create_table | write:catalog-data | Catalog Data Writer |
| load_table | read:catalog-data or write:catalog-data | Catalog Data Reader |
| drop_table | read:catalog-data, write:catalog-data | Catalog Data Writer |
| table_exists | read:catalog-data | Catalog Data Reader |
| rename_table | write:catalog-data | Catalog Data Writer |
| commit_table | read:catalog-data, write:catalog-data | Catalog Data Writer |
| commit_view | read:catalog-data, write:catalog-data | Catalog Data Writer |
| create_view | read:catalog-data, write:catalog-data | Catalog Data Writer |
| drop_view | read:catalog-data, write:catalog-data | Catalog Data Writer |
| view_exists | read:catalog-data | Catalog Data Reader |
| list_views | read:catalog-data | Catalog Data Reader |
| load_view | read:catalog-data | Catalog Data Reader |
| rename_view | write:catalog-data | Catalog Data Writer |
Migration Portal Projects Owner
This role is for users who manage Migration Portal projects in an HM project. They can create, read, update, and delete these projects.
Migration Portal Projects Editor
Users with this role can read and update Migration Portal projects in an HM project.
Migration Portal Projects Viewer
This role provides read-only access to Migration Portal projects in an HM project.
Best practices
- Least privilege: Assign the minimum role that allows the user to perform their tasks. Start with Viewer and escalate only when needed.
- Separate human and machine accounts: Never share human credentials with automation. Create dedicated machine users.
- Audit regularly: Review the activity log monthly to identify unused accounts or over-privileged users.
- Short-lived access keys: Set the shortest practical TTL for access keys. For CI/CD pipelines that run daily, a 7-day key with automated rotation is preferable to a 365-day key.