What This Category of Data Looks Like
Data for this category is primarily sourced from business databases, data dictionaries, desensitization tools, and access approval systems. Updates are completed per batch project or single task. Daily produced documents include desensitization plans, field mapping tables, desensitization verification reports, and approval records. Data fields cover four categories: sensitive field identifiers, desensitization rule codes, execution logs, and approval statuses. Common units are individual record counts, execution durations, and project numbers.
Constraints Imposed by These Characteristics in the Database and Operations Link
Decentralized data sources require integration with multiple types of business databases, data dictionaries, and approval systems. Secure permissions for cross-source data pulling must be configured. The batch update rhythm per project requires databases to support transaction rollback and batch write capabilities. This prevents single project execution failures from affecting overall tasks. Field mapping tables contain sensitive information, so this type of data stored in databases must be encrypted. Approval records require full-link audit trails to be retained. Databases must support precise querying and exporting per project dimension.
How to Set Configurations
These configurations are based on FastGPT v4.17.0.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
MONGODB_URI | mongodb://admin:xxxx@mongodb-host:27017/fastgpt_desensitize?authSource=admin | Used to store desensitization task logs and desensitized dataset metadata, requires specifying the authentication source and dedicated business database |
PG_URL | postgresql://desensitize_user:xxxx@pg-host:5432/desensitize_center?sslmode=require | Stores field mapping tables, approval records, and desensitization verification reports, requires enabling SSL to ensure transmission encryption |
DB_MAX_LINK | 15–25 connections | Adapts to concurrent connection requirements for multiple projects in a single batch, prevents database connection pool exhaustion |
REDIS_URL | redis://:xxxx@redis-host:6379/1 | Caches desensitization rules, temporary task statuses, and sensitive field tags, reduces repeated database query overhead |
| Task Execution Timeout | 3600 seconds | Adapts to full desensitization processes for complex projects, prevents long-running tasks from being forcibly interrupted |
Three Common Mistakes
- Symptom: An internal server error prompt appears when installing database connection plugins, and logs show connection timeout. Cause: Authentication information for
MONGODB_URIorPG_URLis not configured correctly, or the target database has not opened corresponding port access permissions. - Symptom: A MongoDB error prompt appears after service startup, and initialization cannot be completed. Cause: The dedicated business database specified in
MONGODB_URIhas not been created, or the authentication source parameter in the connection string is incorrect. - Symptom: No corresponding records are returned when exporting desensitized data. Cause: Database query filter conditions are not configured per project dimension, or transactions were not committed during batch writing, resulting in data not being persisted.
How to Confirm Configurations Are Correct
- Log into the target database, execute query statements to verify that the databases and tables pointed to by
MONGODB_URIandPG_URLsupport normal read and write operations. - After starting a desensitization task, check whether temporary task status identifiers exist in the Redis cache to confirm that the
REDIS_URLconfiguration takes effect. - After adjusting the
DB_MAX_LINKparameter, observe database connection pool monitoring metrics to confirm that the number of connections does not exceed the set threshold. - After submitting an approval record, query the approval status of the corresponding project in the database to confirm that data has been successfully written.
The values provided are common starting points and should be measured 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.