What Data Looks Like for This Category
Data for this category comes from registered and filed formulas, raw material safety documents, production equipment such as weighing and emulsifying machines, ERP/MES systems, laboratory test records, and packaging material version information. Daily documents include formula process sheets, weighing and feeding records, emulsification or filling logs, cleaning and site clearance records, semi-finished and finished product inspection reports, and label review forms.
Each batch production and line change triggers a data update. Core fields include raw material batch number, emulsification temperature, homogenization duration, finished product batch number, filed formula number, and restricted substance content. Units include degrees Celsius, minutes, batch numbers, and others.
Constraints for Database and Operations
High-frequency data updates triggered by each line change require the database to support low-latency writes and queries. This prevents production process blocking.
Multi-source heterogeneous data includes structured ERP fields, device time-series data, and unstructured inspection reports. Teams must plan indexes and data mapping rules in advance to support this data.
Strong consistency is required between filed formulas and actual production data. Teams must enable transaction-level write verification to avoid risks of inconsistent formulas and filings.
Multi-dimensional data from sensory and stability testing requires full-link traceability associated by batch number. Teams must create a unique index for the batch number.
The traceability requirement for mislabeled product recalls demands tamper-proof copies of full-process data. Operations teams must regularly verify data integrity.
Configuration Settings
The following configurations follow the FastGPT v4.17.0 specification.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
MONGODB_URI | mongodb://user:pass@host:27017/cosmetic_prod?replicaSet=rs0 | Adapts to replica set deployment requirements for cosmetic manufacturing, supports automatic reconnection when the primary node fails over, and meets high availability requirements for multi-source data storage |
PG_URL | postgresql://user:pass@host:5432/mes_data?sslmode=require | Stores structured formula, batch, and inspection data; SSL mode ensures transmission security of filed data |
DB_MAX_LINK | 20–30 | Matches concurrent write and query volumes for each batch production, avoiding production blocking caused by connection pool exhaustion |
REDIS_URL | redis://:pass@host:6379/0?maxmemory-policy=allkeys-lru | Caches frequently queried filed formula and batch mapping information to reduce database load |
| Data Backup Cycle | After each batch production ends + 2:00 AM daily | Covers line changes and daily data updates, ensures full-link data traceability, and meets rapid query requirements during recalls |
| Index Configuration | Create unique indexes for batch number and filed formula number | Accelerates full-link traceability queries by batch, avoids consistency issues caused by duplicate data writes |
Three Common Configuration Mistakes
- Teams may fail to start pgvector or Milvus with configuration files, receiving error messages about incorrect configuration parameters. This occurs when the configuration file does not correctly match the database username with the actual deployed account, and does not include the replica set connection identifier.
- A MongoDB replica set primary node failover may interrupt data synchronization. Teams cannot read the latest production records. This happens when automatic reconnection configuration is not enabled, preventing Change Streams from restoring connections after the primary node switches.
- Private deployment migration may result in empty data fields. Teams cannot associate finished product batch numbers with filed formulas. This occurs when data mapping rules are not verified during migration, and the filed formula number field unique to cosmetic manufacturing is omitted.
How to Confirm Proper Configuration
- Run a batch write test for production data. Check the active connection count of the database connection pool. Confirm the count does not exceed the configured
DB_MAX_LINKlimit. - Trigger a primary node failover simulation. Check MongoDB synchronization status. Confirm Change Streams can automatically restore connections.
- Import a batch of production data. Query full-link records by batch number. Confirm all associated fields return normally.
- Perform a data backup and restore test. Confirm the backup file loads normally. Confirm data integrity meets business requirements.
The values provided are common starting points and should be evaluated against the reader's own samples.