What the data for this category looks like
The data for this category includes documents such as injection molding process cards, raw material drying records, mold histories, first-piece inspections, patrol inspection forms, and material change and machine cleaning records. Data sources include resin batches and property sheets, injection molding machine parameters, mold temperature machines, MES systems, and dimensional and appearance inspection equipment. Data is updated at each production batch, each mold or material change, and each shift patrol inspection node. Fields include resin batch numbers, mold temperature parameters, holding pressure duration, detected dimensional tolerances, and more. Corresponding units are degrees Celsius, seconds, millimeters, and others. The data includes two categories: structured time-series parameters and unstructured documents and images.
What constraints these characteristics impose during the deployment and upgrade phase
This category’s data is multi-source, heterogeneous, and updated at high frequency. During deployment, access formats for multiple data sources must be adapted. During upgrade, version differences in historical production data must be compatible. Since the data includes unstructured documents and images, appropriate storage paths and access permissions must be configured. High-frequency updated time-series data requires optimized database connection configurations to avoid connection exhaustion. Additionally, the format of injection molding process parameters varies slightly across batches. During deployment, unified field mapping rules must be preset. After upgrade, the reading compatibility of historical data must be verified to ensure consistency of traceability logic.
How to set the configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 10–15 | Accommodates high-frequency requests from batch-updated injection molding data, avoids database connection pool exhaustion |
MONGODB_URI | Configure connection string with read-write separation suffix | Stores unstructured documents such as mold histories and patrol inspection forms, supports parallel writing of multi-source data |
PG_URL | Configure connection timeout to 300 seconds | Stores structured time-series data such as injection molding machine parameters and dimensional inspection data, ensures stability of batch queries |
REDIS_URL | Configure memory limit to 2000 MB | Caches frequently read resin property sheets and process card templates, reduces database access pressure |
STORAGE_S3_ENDPOINT | Configure according to enterprise local storage nodes | Stores file-based data such as material change and machine cleaning records and first-piece inspection photos, adapts to existing storage architecture |
Three common mistakes
- Symptom: Imported injection molding process cards lose content after splitting. Cause: Splitting trigger rules are not configured according to the actual fields of the document. Ordinary line breaks within the document are mistakenly used as the only separator, causing long paragraphs to be truncated.
- Symptom: After upgrading to FastGPT v4.17.0, historical mold history data cannot be retrieved. Cause: The connection configuration for
PG_URLis not updated, which is incompatible with the new version’s database permission verification rules. - Symptom: Insufficient permissions prompt appears when generating application API keys. Cause:
ROOT_KEYis not configured as the global management key, and the key generation permission for the corresponding application is not enabled.
How to confirm the configuration is complete
- Log in to the management backend, view the memory usage monitoring of
REDIS_URL, confirm that memory usage meets the configured limit requirements. - Import a complete injection molding process card, check the content integrity after splitting, confirm that no paragraphs are lost or truncated.
- Trigger a batch synchronization of mold history data, confirm that there are no database connection errors and data synchronization is completed.
- Generate an application API key, confirm that the retrieval interface can be called normally and valid results are returned.
The values given are common starting points and should be measured against the reader's own samples.