What the data for this category looks like
The data for this category comes from design BOMs, silicon steel and insulating material purchase ledgers, vacuum drying equipment operation logs, insulating oil test reports, partial discharge and pressure withstand test station records, and product nameplate parameter lists. The data update rhythm follows that each individual product has a complete set of assembly and test process documents, with batch traceability data updated synchronously for each batch of raw materials. The document structure includes dedicated fields such as process node parameters, original test values, and batch codes. There are no universal bulk data templates. Individual product data has a large number of closely related fields, with core fields involving high-voltage electrical dedicated parameters such as insulation distance, water content, and partial discharge peak value.
What constraints these characteristics impose on deployment and upgrade
The dedicated single-product data structure and core high-voltage test fields require that deployment supports custom field mapping and multi-source heterogeneous data access, to adapt to non-standard document formats such as design BOMs and test reports. Following the update rhythm of individual products and batches, deployment must configure an incremental synchronization mechanism to avoid excessive server resource usage from full-volume pulls. The upgrade phase must support gray release to ensure that traceability processes for existing in-production products are not affected. It must also preset verification rules for dedicated fields such as insulation distance and partial discharge peak value, to prevent defect analysis errors caused by data parsing deviations.
How to configure the settings
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
DB_MAX_LINK | 80–120 connections | Maintains a connection pool size suitable for concurrent reading of multiple data sources, as individual products are associated with multiple test and assembly documents |
MONGODB_URI | mongodb://[privileged account]:[password]@[database address]:27017/transformer_bad_analysis?authSource=admin | Stores unstructured document data such as insulating oil test reports and partial discharge test waveforms, using a dedicated instance to isolate business data |
REDIS_URL | redis://:[password]@[cache address]:6379/0 | Caches traceability relationships for individual products and frequently queried test parameters to reduce database access pressure |
STORAGE_S3_ENDPOINT | Enterprise internal object storage access address | Stores large-capacity original test data and assembly record attachments, adapting to the enterprise's existing storage architecture |
FILE_TOKEN_KEY | 16-32 character string generated according to enterprise key specifications | Encrypts access tokens for traceability files, ensuring access security for sensitive power equipment data |
FE_DOMAIN | Enterprise official frontend access domain name | Configures cross-domain access rules to ensure normal communication between frontend pages and backend services |
Three common errors
- Symptom: GPU computing power utilization is low after deployment, and single-core CPU usage consistently reaches 100%. Cause: Dedicated model scheduling rules are not configured for unstructured parsing tasks of high-voltage electrical test data, and general scheduling logic prioritizes CPU resource usage.
- Symptom: After upgrading to FastGPT v4.17.0, an unclickable mask layer appears on the login interface. Cause: Cross-domain rules between
FE_DOMAINand the backend service are not configured correctly, leading to frontend resource loading exceptions and interface rendering failures. - Symptom: The remote WWW service exposes the TRACE request interface, creating a security risk. Cause: TRACE request support for the HTTP service is not disabled during deployment, and the configuration does not comply with the security configuration specifications for power equipment systems.
How to confirm the configuration is complete
- Run a traceability data synchronization test for a single transformer product, verify that associated data such as winding records and pressure withstand test reports can be loaded and displayed normally.
- Check the database connection pool running status, confirm that the current active connection count meets the configuration limit specified by
DB_MAX_LINK. - Access the deployed frontend service address, verify that the interface has no abnormal rendering issues and the login process can be triggered normally.
- Check the HTTP service request configuration, confirm that TRACE requests have been disabled to ensure data access security.
The values provided are common starting points and should be evaluated against the reader's own samples.