What this category of data looks like
This category of data primarily originates from customer requirements, public data sources, collection tools, and in-house sample libraries. Supporting documentation includes collection plans, sample specifications, authorization instructions, and task schedules. Data updates follow single-project cycles, or synchronize the latest samples and schedules on a weekly basis. Core daily processing steps include sample scope definition, field specification confirmation, and collection legality verification. Data deviations or compliance issues may result in sample unavailability, infringement risks, or project rework.
Constraints imposed on permission and data isolation by these characteristics
Since collection task data sources involve customer-specific requirements and authorized data sources, samples and documents for different projects have clear ownership boundaries. Permission isolation must therefore strictly match project and tenant dimensions. Collection data includes authorization instructions and sensitive field configurations, so access to compliance-related content must be restricted for unauthorized personnel. Data updates follow project or weekly cycles, so permission rules must support dynamic adaptation to new collection tasks and data sources. This prevents data leaks or permission chaos caused by delayed configuration.
How to set configurations
| Configuration Item | Recommended Approach | Rationale |
|---|---|---|
| Team tenant binding | Bind teamId with project-specific ID | Collection tasks divide data sources by project, matching exclusive ownership boundaries |
| Member permission assignment | Bind tmbId with the tenant ID of the member's affiliation | Collection data for different tenants must be fully isolated to prevent cross-tenant access |
| Sensitive field filtering | Filter by matching field names | Collection data includes authorization-restricted fields, so access must be limited for unauthorized personnel |
| Storage directory isolation | Allocate a 1000 MB exclusive storage directory | Collection sample data volume grows with projects; exclusive directories prevent data mixing |
| Model call association | Associate call records by teamId | Match the team's model cost statistics requirements, enabling billing by team dimension |
| Task visibility scope | Limit to tasks corresponding to the current teamId | Prevent unauthorized personnel from viewing cross-project collection plans and schedules |
Three common configuration mistakes
- Team members cannot view cross-project collection task lists. The cause is failure to correctly bind
teamId, resulting in unrestricted permission scope and misconfiguration for global access. - Exported collection sample data includes unauthorized public data source content. The cause is failure to configure tenant isolation rules corresponding to
tmbId, leading to unrestricted cross-tenant data access. - Costs generated by model calls from collection tasks of different teams cannot be counted separately. The cause is failure to associate model call records by binding
teamIdper team, leading to mixed cost statistics.
How to confirm the configuration is complete
- Access FastGPT v4.17.0, navigate to the collection task management page, and verify that only tasks corresponding to the current
teamIdare visible. - Use an unauthorized tenant account to attempt access, and confirm that collection plans and sample specification documents of other tenants cannot be viewed.
- Export collection sample data, and verify that it only includes data sources within the current project's authorization scope, with no unauthorized data.
- View model call logs, and confirm that only call records associated with the current team's
teamIdare counted separately.
The values provided here are common starting points, and should be evaluated against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-10-19.