What the data for this category looks like
This category's data originates from production execution systems, testing equipment, laboratory systems, process parameter documents, customer standards, and historical defect databases. Daily documents include control plans, patrol inspection records, SPC reports, finished product inspection reports, nonconforming product review records, release orders, and quality monthly reports. Core fields include batch identifiers, inspection item parameters, judgment results, defect types, and more. Updates are collected continuously alongside production batches and shifts. The data exists in both structured report and unstructured technical document formats.
What constraints these characteristics impose on deployment and upgrade workflows
Data for this category comes from multiple dispersed sources and is updated frequently. During deployment, configure multi-source data synchronization links to ensure real-time access to data from production, testing, laboratory, and other systems. Documents have both structured and unstructured formats. During deployment, configure separate parsing rules to adapt to different formats of technical documents and reports. Data is updated continuously with production batches. During upgrades, verify the incremental index update logic to avoid inconsistent judgment standards for new and old data caused by version changes. Additionally, data involves finished product release permissions. During deployment, bind access control for corresponding roles. During upgrades, ensure permission configurations are not overwritten to prevent accidental changes to the release judgment process. Since data is associated with production batches and quality judgments, reserve a data migration window during upgrades to avoid disrupting ongoing in-process product release workflows.
How to set configurations
This configuration is based on FastGPT v4.17.0, and adapts to knowledge base deployment and upgrade requirements for process quality and finished product release scenarios.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
MONGODB_URI | mongodb://qc_user:qc_pass@db-host:27017/fastgpt_qc | Stores structured quality data, requires a dedicated database and access credentials |
REDIS_URL | redis://:redis_pass@redis-host:6379/1 | Caches frequently queried quality documents and permission configurations, uses database 1 to isolate business data |
STORAGE_S3_ENDPOINT | http://storage.example.com:9000 | Stores unstructured process procedures and inspection reports, matches the address of a self-built object storage service |
DB_MAX_LINK | 25–35 connections | Adapts to the high-frequency quality data query requirements of this category, avoids connection exhaustion |
FILE_TOKEN_KEY | 32-digit random encryption string | Encrypts knowledge base file upload tokens, ensures access security for quality documents |
Three common errors
- Issue: After upgrading to a new version, knowledge base queries return empty results or historical quality data becomes invalid. Cause: Incremental vector database data migration was not performed. Only the application code was updated, and the index structure was not synchronized. This causes a mismatch between vector dimensions or index rules for new and old data.
- Issue: Calling the index interface returns a 400 status code with no response body. Cause: After upgrading, the access key or configuration parameters for the index service were not updated, leading to permission verification failure. The new version's index request format was not adapted either.
- Issue: FastGPT cannot upload knowledge base files after Docker deployment. Cause:
STORAGE_S3_ENDPOINTor the local storage path was not configured correctly, interrupting the file upload link. The correct file upload token key was not set either.
How to confirm the configuration is complete
- Log in to the FastGPT admin backend, check the connectivity test results for database configuration items, and confirm they show normal status.
- Upload a typical finished product inspection report file, verify that the file upload and parsing process completes normally with no error prompts.
- Trigger an incremental data synchronization, check that the index service can normally update the vector index of quality data, with no abnormal logs.
- Configure a test release judgment query, input historical inspection data, and verify that the returned results match the preset standards.
The values provided are common starting points and should be evaluated against local deployment samples.