What this category of data looks like
Data for this category covers multiple document types across the full heat treatment workflow. These include heat treatment process cards, charging charts, furnace temperature curves, hardness and metallographic reports, equipment inspection forms, and exception handling sheets. Data sources include material grades, customer technical requirements, temperature control instruments, furnace temperature uniformity testing equipment, hardness testers and metallographic equipment, and workshop MES systems. Data updates occur per furnace charge or batch. Documents generated for a single furnace charge include fields with clear units: furnace temperature (℃), holding time (hours), hardness value (HV), metallographic structure grade, and others. Document structures vary significantly: some are structured instrument data, while others are unstructured reports and images.
What constraints these characteristics impose on deployment and upgrade
This category of data is updated frequently, is multi-source and heterogeneous, and is tied to strict process parameters. Deployment must support high-concurrency, small-batch write requests, plus storage and retrieval of multi-format files. Some heat treatment workshops use offline servers, so upgrade steps must be compatible with offline environments and avoid relying on public network resources. The specificity of data fields and units requires configuration steps that support custom metadata mapping. This ensures that inspection and audit workflows can accurately match business fields. Since inspection results directly relate to product performance, data validation rules must be configured after deployment. This prevents incorrect data from entering audit workflows.
How to set up the configuration
The following configurations are optimized for FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 connections | Adapts to concurrent write and query requests for multiple documents per furnace charge, and prevents database connection limit breaches |
REDIS_URL | redis://:custom password@intranet IP:6379/0 | Caches session data and temporary inspection documents, and supports batch processing workflows |
STORAGE_S3_ENDPOINT | Intranet MinIO address or public OSS endpoint | Stores multi-format large files such as furnace temperature curves and metallographic reports, and supports offline or cloud deployment scenarios |
FE_DOMAIN | Heat treatment workshop internal domain or filed public domain | Configures the HTTPS access path, and supports local or remote operation and maintenance needs |
MONGODB_URI | mongodb://username:password@intranet IP:27017/fastgpt?authSource=admin | Stores inspection records and audit logs, and supports high-frequency write and query operations |
FILE_TOKEN_KEY | Randomly generated 32-bit string | Encrypts file access tokens, and ensures secure access to inspection documents |
Three common mistakes
- Accessing the FastGPT interface shows an insecure certificate warning, and HTTPS connections cannot be established. The cause is failure to mount the certificate generated by mkcert to the corresponding service container, and failure to associate the certificate path with the
FE_DOMAINconfiguration. - The AI proxy reports an error
get tiktoken dial tcp lookupduring offline server deployment. The cause is that the offline environment has no local tokenizer cache configured, or no offline model path is specified, causing the program to attempt to obtain resources over the public network and fail. - Submitting an audit workflow returns no response with a 500 status code after public cloud deployment. The cause is incorrect permission configuration in
MONGODB_URIorREDIS_URL, or failure to open intranet access permissions for the corresponding ports.
How to confirm the configuration is valid
- Log in to the system monitoring page in the FastGPT backend, check the database connection count, and confirm that no
DB_MAX_LINKconnection limit breach alert is triggered. - Upload a heat treatment furnace temperature curve PDF file, verify that the file can be stored and previewed normally, and confirm that the
STORAGE_S3_ENDPOINTconfiguration takes effect. - Initiate a single furnace charge inspection record audit workflow, check the session logs, and confirm that there are no exceptions in cache read and write operations for
REDIS_URL. - Access the address corresponding to
FE_DOMAIN, confirm that the browser shows a secure HTTPS connection with no certificate warnings.
The values provided here are common starting points, and should be evaluated against the reader’s own samples.