What this category of data looks like
Data for this category comes from slab or billet batches, heating furnaces, mill automation control systems, online thickness and width measuring devices, surface inspection equipment, and MES systems of hot rolling, cold rolling, section steel, and steel pipe enterprises. Daily generated documents include rolling plans, heating schedules, rolling regulations, flatness and thickness records, surface quality reports, and quality certificates. Data updates in real time per coil or individual steel piece. Document fields cover process parameters, quality test results, and batch identifiers. Associated units include degrees Celsius, percentage, millimeters, and others.
What constraints these characteristics impose during deployment and upgrade
Data for this category updates in real time per coil, and sources are scattered. During deployment, support must be provided for real-time data access from multiple device and system types, to avoid performance loss from full synchronization. Coil-level traceability requires high data association accuracy. Configuration must ensure unique binding of batch identifiers. Diverse daily document types require adaptation for parsing and storage of different formats such as rolling regulations and quality reports. During upgrade, continuity of existing production systems must be maintained, to avoid disrupting real-time process data flow due to configuration changes.
How to set configurations
These configurations apply to FastGPT v4.17.0:
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 20–30 connections | Data for this category updates in real time per coil. High-frequency data writes and queries require a reasonable database connection pool size |
REDIS_URL | primary node + 1 replica node | Real-time process and quality data require low-latency caching to enable fast retrieval and association of coil-level data |
STORAGE_S3_ENDPOINT | industrial-grade distributed storage endpoint | Documents and quality data in this category have large volume, requiring stable distributed storage to support batch file storage |
| Single file upload limit | 1000 MB | A single rolling regulation or quality certificate document may contain multiple pages of test data, so large file upload support is required |
FILE_TOKEN_KEY | Calibrated via actual testing | Need to ensure the security of document access tokens, and adapt to internal enterprise permission control rules |
ROOT_KEY | Calibrated via actual testing | Ensure permission security for internal system interface calls, and adapt to the permission system for enterprise-level deployment |
Three common mistakes
- Phenomenon: The file processing plugin cannot be found in the interface, with no corresponding option in the plugin list. Cause: The corresponding version of the plugin package was not installed during deployment, or the FastGPT version does not match the plugin version.
- Phenomenon: An error occurs when creating a knowledge base using a URL. The backend log shows
cannot fetch internal url. Cause: The access address of the internal service was not configured correctly after upgrade, causing the system to fail to pull internal resources normally. - Phenomenon: An error occurs when running the docker run command to deploy the image parsing component. Cause: The GPU mapping parameter was not specified correctly, or the local GPU driver version is incompatible with the image version, causing the component to fail to start normally.
How to confirm the configuration is complete
- Run the database connection test script to verify that the configured connection pool size can adapt to current business concurrency. The threshold should be calibrated based on actual production data flow volume.
- Upload a rolling regulation document containing multiple pages of test data to confirm that the file upload and storage process works normally, with no errors related to size limits.
- Check the system backend logs to confirm that the connection status between the cache node and storage endpoint is normal, with no timeout or permission error prompts.
- Test the process of creating a knowledge base using an internal data source URL to confirm that the resource pulling function works normally, with no internal access errors.
The values given are common starting points and should be measured against the reader's own samples.