What this category of data looks like
This category of data includes foundry process cards, batching lists, heat records, mold history records, first-piece inspection records, material test reports, and nonconforming product review forms. Data comes from ERP material master data, melting furnaces and thermometers, spectrometers, MES heat records, mold management systems, and customer drawings and technical specifications. Updates run per heat, per batch, or daily. Fields include alloy batching ratio, melting temperature, pouring parameters, defect type, heat and mold numbers, and others. Units include degrees Celsius, kilograms, millimeters, and additional relevant units.
Constraints for Deployment and Upgrade
The multi-source heterogeneous nature of this data category requires adapting to different document parsing formats during deployment. It supports mixed processing of structured tables (like batching lists) and unstructured text (like review forms). Frequently updated heat data needs a stable real-time data synchronization link during deployment. This avoids audit delays from synchronization latency. Precise verification of multiple fields requires retaining original field mapping rules during upgrades. This prevents configured parameter matching logic from breaking during version iterations. Nonconforming product data for key products needs a complete traceability chain. Deployment must ensure data storage compliance and traceability.
How to Set Configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 10–15 connections | Adapt to concurrent connection requirements for multi-source data synchronization, avoid exhausting database connections |
MONGODB_URI | Enter the dedicated MongoDB connection string for foundry operations | Isolate foundry scenario data from other business data to ensure data security |
REDIS_URL | Configure the connection address for the local Redis instance | Cache frequently queried heat records and parameter data to improve audit response speed |
FE_DOMAIN | http://your-casting-deploy-domain | Support cross-device access for login-free link generation, meet workshop inspection usage requirements |
STORAGE_S3_ENDPOINT | `http://internal-s3-endpoint:9000 | Store unstructured files such as heat photos and spectral test reports, uniformly manage business attachments |
Three Common Mistakes
- Symptom: A "Connection failed" error appears when opening the workflow page after Docker deployment. Cause: Incorrect configuration of
REDIS_URLorMONGODB_URI. This prevents the workflow from accessing cached or database process data. - Symptom: Saved custom model configurations are replaced with other models when re-entering the application after upgrading to FastGPT v4.17.0. Cause: The upgrade script does not retain the original model mapping configuration. This overwrites configuration items.
- Symptom: A "Username or password incorrect" prompt appears when logging into the system after deployment is complete. Cause: Failure to execute the initialization command correctly, or formatting errors in the
ROOT_KEYenvironment variable configuration. This prevents the administrator account from generating properly.
How to Confirm Configuration Success
- Access the configured
FE_DOMAINaddress. Confirm the frontend page loads normally with no resource loading failure prompts. - Run a database connection test. Verify that the connection count configured by
DB_MAX_LINKmeets current business concurrent access requirements. - Upload a foundry batching list document. Confirm parsed fields correctly map to preset audit rules.
- Generate a login-free link. Open it on another device. Confirm the application loads normally without login required.
The values provided are common starting points and should be evaluated against the reader's own samples.