TVL Managed Superset

Advanced RBAC on Apache Superset: Role Design 2026

Design advanced RBAC on Apache Superset: business roles, fine permissions, separation of concerns.

Beyond the 5 native roles, Apache Superset allows building advanced RBAC with business roles, fine permissions, and separation of concerns. This guide explains how to design RBAC that scales for 100+ users in 2026.

1. Native role limits

Admin/Alpha/Gamma roles are insufficient at scale:

  • Too broad: Alpha gives too many rights to a business analyst;
  • Not business-oriented: no pre-made "Marketing" or "Finance" role;
  • Not hierarchical: everything is flat;
  • No dynamic RBAC per dataset.

If you want pre-designed RBAC, TVL Managed Superset offers business RBAC templates on Pro+ instances.

2. Recommended pattern

Mix technical roles and business roles:

Technical role+ Business role
Gamma+ Marketing → sees marketing datasets
Gamma+ Finance → sees finance datasets
Alpha+ Data Eng → all datasets
Admin(no complement)

3. Create a "Marketing" business role

  1. UI → Settings → List Roles → + Role;
  2. Name: Marketing;
  3. Permissions:
    • can read on Database (warehouse);
    • schema_access on marketing.*;
    • datasource_access on ad-hoc datasets.
  4. Save.

4. Available fine permissions

PermissionEffect
can_read on DashboardView dashboards
can_write on DashboardCreate/edit dashboards
can_export on DashboardPDF/PNG export
can_sql_json on SupersetSQL Lab access
database_access[mydb]Specific DB connection
schema_access[mydb.public]Specific schema
datasource_access[ds_id]Specific dataset
menu_access on SettingsSettings menu access

This configuration is applied by default on TVL Managed Superset, which follows community best practices.

5. Separation of concerns

  • Admin: 2-3 people max;
  • Data Engineering: creates datasets, models in dbt;
  • Data Analysts: creates charts/dashboards from ready datasets;
  • Business: consumes dashboards, can create simple slices.

6. RBAC on sensitive datasets

For HR, finance, salary data:

  • Separate datasets with restricted datasource_access;
  • RLS as complement (cf. RLS);
  • Audit log viewable only by compliance.

7. SSO mapping

Synchronize roles with the IdP:

AUTH_ROLES_MAPPING = {
    "ad_marketing": ["Gamma", "Marketing"],
    "ad_finance":   ["Gamma", "Finance"],
    "ad_data_eng":  ["Alpha", "Data Engineering"],
    "ad_admins":    ["Admin"],
}
AUTH_ROLES_SYNC_AT_LOGIN = True

8. Audit and governance

  • Quarterly review of assigned roles;
  • Auto deletion of inactive accounts > 90 days;
  • Audit log of role changes;
  • Documentation of RBAC rules in Confluence/Notion.

9. Common pitfalls

  • Too many roles (50+) → unmanageable;
  • Permissions accumulated over time without review;
  • Public role forgotten → leak;
  • RBAC without RLS → a user sees all tenants via SQL Lab;
  • No naming convention making mapping incomprehensible.

10. Conclusion

A well-designed Apache Superset RBAC combines technical roles (Admin/Alpha/Gamma) and business roles (Marketing/Finance/Sales) with automatic SSO mapping. At scale (100+ users), it's the only way to keep control without saturating admins.

Want the benefits of Apache Superset without the friction of installation and maintenance? Deploy your instance in 3 clicks with TVL Managed Superset, hosted in Europe (OVHcloud, Roubaix, France).

For more: roles and permissions, RLS, audit trail.