The authorization-system microservice provides centralized authorization services for the FlowX.AI platform, managing workspaces, users, groups, roles, and permissions. It works alongside SpiceDB to deliver fine-grained access control and supports the Workspaces feature.

​
Database configuration

The authorization-system must use a dedicated PostgreSQL database. Do not share with other FlowX.AI services.
Environment Variables
SPRING_DATASOURCE_URL=jdbc:postgresql://postgresql:5432/authorization_system
SPRING_DATASOURCE_USERNAME=flowx
SPRING_DATASOURCE_PASSWORD=<password>  # Use Kubernetes Secret
Requirements:
  • Database user needs full access to authorization_system database
  • PostgreSQL must be available before service startup

​
CAS client library configuration

The authorization-system uses the CAS client library to communicate with SpiceDB for ACL operations.
Environment Variables
FLOWX_LIB_CASCLIENT_SPICEDB_HOST=spicedb
FLOWX_LIB_CASCLIENT_SPICEDB_PORT=50051
FLOWX_LIB_CASCLIENT_SPICEDB_TOKEN=<spicedb-token>  # Use Kubernetes Secret
The SpiceDB token is the SPICEDB_GRPC_PRESHARED_KEY value from your SpiceDB configuration.
Configuration Parameters:
  • SpiceDB Host: Service hostname (typically spicedb)
  • SpiceDB Port: gRPC port (standard: 50051)
  • SpiceDB Token: Authentication token for SpiceDB access

​
OAuth2/Keycloak configuration

Environment Variables
SECURITY_OAUTH2_BASE_SERVER_URL=https://auth.yourcompany.com/auth
SECURITY_OAUTH2_REALM=flowx
SECURITY_OAUTH2_CLIENT_CLIENT_ID=flowx-platform-authorize
SECURITY_OAUTH2_CLIENT_CLIENT_SECRET=<oauth-secret>  # Use Kubernetes Secret
SECURITY_OAUTH2_SERVICE_ACCOUNT_ADMIN_CLIENT_SECRET=<service-account-secret>  # Use Kubernetes Secret

OAuth2 Client

Must be configured in Keycloak with appropriate scopes for platform access

Service Account

Requires admin privileges in Keycloak for user management operations

​
Redis configuration

Environment Variables
SPRING_REDIS_HOST=redis-master
SPRING_REDIS_PASSWORD=<redis-password>  # Use Kubernetes Secret

​
Management

Environment Variables
MANAGEMENT_SERVER_PORT=8081

​
Organization admin bootstrap

The authorization-system uses a fallback mechanism to create the first organization administrator when no admin users exist.
Set SPRING_LIQUIBASE_PARAMETERS_DEFAULTORGADMINUSERNAME (default: admin@flowx.ai) Process:
  • System searches for this username in Keycloak
  • Copies the user’s sub_id (subject ID) to authorization-system database
  • Grants organization admin privileges automatically

​
Fallback method

Set SPRING_LIQUIBASE_PARAMETERS_DEFAULTORGADMINUSERSUBJECTID with a specific Keycloak subject ID Process:
  • Creates user directly in authorization-system database
  • Assigns organization admin roles
  • Used when username method fails or is set to null

​
Error handling

If incorrect subject_id is provided:
  • Login will fail
  • No org-admin privileges granted
  • Manual database correction required

​
Customer-specific variables

Required Customization: These variables must be updated for each deployment environment.
  • SECURITY_OAUTH2_BASE_SERVER_URL - Your Keycloak server URL
  • SECURITY_OAUTH2_REALM - Your Keycloak realm name
  • SECURITY_OAUTH2_CLIENT_CLIENT_ID - Your OAuth2 client identifier
  • SPRING_DATASOURCE_URL - Your PostgreSQL connection details
  • Service hostnames - Update to match your Kubernetes service names

​
Secrets management

Security: Always use Kubernetes Secrets for sensitive configuration values.
Required Kubernetes Secrets:
  • SPRING_DATASOURCE_PASSWORD
  • FLOWX_LIB_CASCLIENT_SPICEDB_TOKEN
  • SPRING_REDIS_PASSWORD
  • SECURITY_OAUTH2_CLIENT_CLIENT_SECRET
  • SECURITY_OAUTH2_SERVICE_ACCOUNT_ADMIN_CLIENT_SECRET

​
Deployment prerequisites

Infrastructure

  • PostgreSQL with authorization_system database
  • SpiceDB with authentication configured
  • Redis for caching

Identity & Access

  • Keycloak with configured realm
  • OAuth2 clients created
  • Admin user exists in Keycloak

​
Architecture notes

Database Access Control: Only authorization-system has direct write access to the CAS PostgreSQL database. Other services communicate via REST APIs only.
SpiceDB Integration: Uses PostgreSQL as backend storage and communicates via gRPC through the CAS client library.