What the data for this category looks like
Core data for this category comes from formulation systems, raw material batch records, internal mixers, Mooney viscometers and rheometers, vulcanizers, and laboratory testing equipment. Daily documents include formulation sheets, weighing records, mixing curves, rubber compound quick test reports, vulcanization process cards, mold usage records, and finished product performance reports.
Data updates occur per truckload of rubber, per batch of raw materials, and per mold production operation. Single data entries include fields such as feeding amount (kilograms), mixing temperature (degrees Celsius), vulcanization time (seconds), and finished product tensile strength (megapascals). Supported formats include structured tables, time-series curves, and text records.
Constraints imposed by these characteristics during deployment and upgrade
High-frequency single-entry data updates require database connection pool configurations adapted for high-concurrency writes, to avoid service interruptions caused by exhausted connections. Multi-source heterogeneous data sources require unified storage and permission management to ensure that mixing curves and laboratory reports uploaded from different devices can be retrieved uniformly.
The per-mold update rhythm requires cache configurations adapted for frequently read process parameters, to reduce access pressure on the underlying database. Additionally, sensitive documents such as formulation sheets and performance reports require access permission controls to prevent leakage of core process data.
During deployment, compatibility with ARM architecture devices must be considered, along with resource constraints for edge deployment scenarios such as soft routers.
How to set the configurations
The following configurations are based on FastGPT v4.17.0
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 | Adapts to high-frequency data writes per truckload of rubber and per mold production, avoiding exhaustion of the database connection pool |
MONGODB_URI | mongodb://[username]:[password]@[host]:27017/fastgpt?authSource=admin | Stores structured documents such as formulation sheets and performance reports, supporting persistent storage of multi-source data |
REDIS_URL | redis://:[password]@[host]:6379/0 | Caches frequently read mixing curves and vulcanization process cards, reducing repeated query pressure on the database |
STORAGE_S3_ENDPOINT | Fill in according to the local object storage address | Uniformly stores binary files such as mixing curves and mold records uploaded by internal mixers, simplifying file management |
FILE_TOKEN_KEY | Randomly generated 32-bit string | Protects access permissions for sensitive documents such as uploaded formulation sheets and performance reports, preventing unauthorized access |
FE_DOMAIN | http://[local machine IP]:3000 (local deployment) or the bound official domain name | Configures the frontend access address to ensure normal loading of frontend pages and resource calls |
Three common misconfigurations
- Symptom: After deployment, the frontend page cannot be opened. Port detection shows the port is open, but page loading fails. Cause: The
FE_DOMAINconfiguration is not bound to the correct local IP or port, leading to incorrect frontend resource paths and failure to load page resources. - Symptom: When connecting an external vllm model, oneapi displays a connection refused error. Cause: Incorrect
MONGODB_URIorREDIS_URLconfiguration, causing the FastGPT service to fail to initialize properly, or incorrect port mapping configuration blocking connection requests from the model service. - Symptom: When deploying via Docker, the downloaded yml file only contains a single line of 404 content. Cause: The deployment file download link is invalid or the network is interrupted, leading to incomplete file download. Obtain the official deployment file again.
How to confirm successful configuration
- Execute the
docker pscommand to confirm that all FastGPT-related containers are in a running state. - Access the address configured in
FE_DOMAINto confirm that the frontend page loads normally with no resource loading failure prompts. - Connect to the database configured in
MONGODB_URIto verify that test formulation sheet data can be written, confirming normal storage functionality. - Call the FastGPT built-in model test interface to confirm that external vector models and language model APIs can connect normally and return results.
The values provided here are common starting points and should be evaluated against appropriate samples.