What the data for this category looks like
Data sources for this category include wafer test maps, MES production management systems, bonding and molding equipment logs, data collected by ATE test machines, X-ray/SAM inspection results, and customer specifications. Daily generated documents cover packaging process cards, bonding and molding parameter tables, incoming wafer maps, test program version files, yield reports, and failure analysis reports. Data update frequencies cover each production batch, each chip test cycle, and daily summary cycles. Documents include dedicated fields such as wafer ID, test bin value, parameter thresholds, and batch number. Units involve electronic manufacturing-specific metrology standards including voltage, time, and yield percentage.
What constraints these characteristics impose on the deployment and upgrade phase
The data sources for this category cover industrial equipment, production systems, and customer documents. The high-frequency update feature requires the deployment phase to support low-latency data access and parsing. Multiple types of documents include dedicated fields and units, so document parsing rules adapted to electronic manufacturing scenarios must be configured during deployment. The time sensitivity of production links requires that core document processing processes not be interrupted during upgrades, so a rolling upgrade method must be used. At the same time, cross-system docking requirements require configuration items to be compatible with interface specifications of industrial systems such as MES and ATE test machines, to avoid production schedule impacts from data synchronization interruptions.
How to set the configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 connections | Adapts to the high-frequency data write and query requirements of this category, matches the industrial-grade database connection pool limit of FastGPT v4.17.0 |
STORAGE_S3_ENDPOINT | Consistent with the intranet address of the local MinIO instance, e.g. http://192.168.1.100:9000 | Document storage for this category requires adaptation to low-latency access in industrial intranets, avoids delays from public network transmission, and ensures production data synchronization efficiency |
REDIS_URL | Intranet Redis instance address, e.g. redis://:password@192.168.1.101:6379/0 | Used for high-frequency caching of temporary data such as wafer batches and test data, reducing query pressure on the core database |
MONGODB_URI | Configure multiple addresses in a concatenated format per cluster node, e.g. mongodb://user:pass@host1:27017,host2:27017/admin?replicaSet=rs0 | Ensures high availability for high-concurrency document storage of this category, avoids document cross-review process impacts from single-node failures |
FE_DOMAIN | Production environment intranet domain name, e.g. fastgpt-seal.test.com | Only allows access from the factory intranet, ensuring access security for industrial document data |
Three common mistakes
- Symptom: Slow download speed when pulling the
fastgpt-minioimage, with download timeout errors displayed in the interface. Cause: No domestic mirror acceleration source is configured, and pulling the official image directly leads to excessive cross-network transmission delays. - Symptom: Unable to add image datasets after private deployment, with the image upload button missing in the knowledge base page. Cause:
STORAGE_S3_ENDPOINTand associated storage keys are not correctly configured, causing the file storage module to fail to load properly. - Symptom: Low matching accuracy of cross-review results, with incoming wafer maps unable to be correctly associated with batch data. Cause: Document indexing rules are not configured based on the dedicated fields of this category (such as wafer ID and test bin value), leading to semantic matching deviations.
How to confirm the configuration is complete
- Run the database connection test script to verify that configuration items such as
MONGODB_URIandREDIS_URLcan establish normal connections, and confirm that the connection pool quantity meets theDB_MAX_LINKconfiguration range. - Upload two typical documents: packaging process card and wafer test map, check that files can be stored normally and retrieved in the knowledge base, to verify that the
STORAGE_S3_ENDPOINTconfiguration takes effect. - Access the intranet address corresponding to
FE_DOMAIN, confirm that the FastGPT frontend page loads normally, and that the tag function of the knowledge base module can be enabled normally. - Simulate high-frequency data writes, observe that the database connection count does not exceed the
DB_MAX_LINKconfiguration value, and confirm that system load meets expectations.
The values provided are common starting points and should be evaluated against the reader's own samples.