What data for this category looks like
The data for this category mainly comes from CAD/CAM software, order BOM systems, sheet metal inventory ledgers, laser cutting machine and bending machine device logs, and workshop MES systems. Daily generated documents include 3D part unfolding drawings, CNC cutting programs, bending process cards, nesting diagrams, first-piece inspection records, and process transfer cards. Data updates synchronize with order delivery, daily inventory checks, and each production changeover. Document fields include bending compensation coefficients, hole position coordinates, sheet metal specifications, bending sequences, process transfer nodes, and some documents are associated with sheet metal material and thickness parameters.
What constraints these characteristics impose during deployment and upgrade
The data for this category is multi-source, heterogeneous, and updated frequently. During deployment, adapt to multiple data access methods such as CAD/CAM export formats and MES interface protocols, and pre-configure corresponding parsing rules. High-frequency order and changeover data updates require deployed services to have stable low-latency read and write capabilities. During upgrades, optimize database connection pool parameters. Multiple types of process documents include professional engineering fields such as bending coefficients and hole position coordinates. During deployment, configure vector indexes that support engineering attributes. Fluctuations in document volume caused by multi-variety and small-batch production require deployed storage resources to have elastic adaptation capabilities, to avoid service freezes caused by sudden increases in data volume. In addition, cross-module association of core process data requires ensuring consistency of data links during deployment, to avoid process errors caused by synchronization delays.
How to configure the settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 10–15 | Adapt to the concurrent read and write requirements of multi-source devices (laser cutting machines, bending machines, MES) for this category, and support real-time interaction of workshop production data |
MONGODB_URI | mongodb://username:password@host:27017/fastgpt?authSource=admin | Store unstructured data such as process documents and process transfer records. Configure a connection address with permission verification to comply with the data storage specifications of FastGPT v4.17.0 |
PG_URL | postgresql://username:password@host:5432/fastgpt | Store structured business data such as order BOMs and sheet metal inventory, to support complex query requirements for relational data |
REDIS_URL | redis://:password@host:6379/0 | Cache frequently accessed data such as bending coefficients and nesting rules, to reduce repeated read and write pressure on the database |
STORAGE_S3_ENDPOINT | Fill in according to the local object storage address, such as http://192.168.1.100:9000 | Store large-volume engineering files such as unfolding drawings and CNC programs, to adapt to private cloud or on-premises storage deployment scenarios |
FILE_TOKEN_KEY | Random string calibrated through actual testing | Generate access tokens for engineering files, to ensure the security of process document downloads |
Three common configuration mistakes
- Container restarts continuously and the frontend cannot be accessed. The symptom is that FastGPT v4.17.0 container logs prompt MongoDB connection authentication failure or timeout. The cause is that the username, password, or database name in
MONGODB_URIare not configured correctly, causing the service to fail to initialize the core data storage. - Engineering document parsing fails. The symptom is that uploaded sheet metal unfolding drawings or nesting diagrams cannot generate vector indexes. The cause is that a parsing adapter adapted to CAD formats is not configured, causing professional fields such as bending coefficients and hole position coordinates to fail to be extracted correctly.
- Slow service response. The symptom is timeouts when querying bending process cards. The cause is that the value of
DB_MAX_LINKis too low, failing to meet the read and write requirements of multiple devices submitting production data concurrently.
How to verify successful configuration
- Log in to the system settings page of the FastGPT backend, check whether the formats and parameters of database connection configurations such as
MONGODB_URIandPG_URLmatch the locally deployed databases. - Upload a standard sheet metal unfolding drawing document, verify that it can be successfully parsed and generate a corresponding vector index, to confirm that the file storage and parsing links are normal.
- Start multiple simulated concurrent requests, observe the service running status, and confirm that the value of
DB_MAX_LINKcan support concurrent access requirements. - View the real-time container log output, confirm that there are no error messages such as database connection failures or authentication errors.
The values given are common starting points and should be measured against the reader's own samples.