What Data Looks Like for This Category
The data for this category comes from formula and process libraries, raw material batch records, reactor DCS data, weighing systems, LIMS systems, and warehouse systems. Multiple types of documents are generated daily, including batch production instructions, feeding review sheets, reaction process records, cleaning records, test reports, deviation and change orders, etc. Data updates follow the rhythm of each batch production or each key process step. Scenarios with multi-product, multi-step batch switching require prioritizing formula confidentiality and the accuracy of batch change cleaning confirmation.
Constraints on Deployment and Upgrade
Connecting multiple heterogeneous data sources for this category requires completing connection configurations for multiple system types during deployment. Frequent batch-level data updates require ensuring the incremental synchronization mechanism is compatible during upgrades. Scenarios with multiple document types and batch switching require configuring index rules adapted for multi-field validation during deployment. Formula confidentiality requirements require configuring permission isolation parameters during storage and access. The accuracy of batch change cleaning confirmation depends on precise document association, which places clear requirements on database connection stability. The rationality of connection pool parameters must be verified during deployment.
How to Configure Settings
The following configurations apply to FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 个 | Adapt to multi-batch concurrent document review requests, avoid connection pool exhaustion |
MONGODB_URI | Configure according to the enterprise MongoDB cluster address, add read-write separation parameters | Store unstructured process records and deviation and change orders, support high-concurrency queries |
PG_URL | postgresql://用户名:密码@地址:5432/数据库名 | Store structured feeding data and batch metadata, ensure transaction consistency |
REDIS_URL | redis://:访问密码@容器内地址:6379/0 | Cache frequently accessed formula summaries and cleaning validation rules, reduce database load |
STORAGE_S3_ENDPOINT | Configure as the endpoint address of the enterprise internal object storage | Store encrypted process documents and batch data, ensure formula confidentiality |
ROOT_KEY | 32 字符的随机字符串 | Configure system initial administrator permissions, isolate unauthorized access |
Three Common Misconfigurations
- Phenomenon: Docker deployment logs show database connection failure, and login cannot be completed. Cause: The access address, account password, or container network for
PG_URLorMONGODB_URIis not correctly configured, and the container network does not connect to the database service. - Phenomenon: When deploying locally on Windows, the specified port cannot be accessed after the container starts. Cause: The local firewall does not open the corresponding port, or the front-end access address configuration does not match the local network environment.
- Phenomenon: Batch process document indexing fails after upgrade, and no matching results are returned. Cause:
DB_MAX_LINKis not adjusted to adapt to the concurrent requests after upgrade, causing the database connection pool to be exhausted and unable to read batch metadata.
How to Verify Successful Configuration
- Run the Docker container log viewing command for the corresponding environment, confirm that there are no error messages about database connection failure or port binding failure.
- Access the configured
FE_DOMAINaddress, verify that the front-end page loads normally and initial administrator account configuration can be completed. - Upload a single batch production instruction document, trigger the indexing process, confirm that the indexing task completes normally and generates a corresponding retrieval entry.
- Check the storage service corresponding to
STORAGE_S3_ENDPOINT, confirm that the uploaded process documents are stored as required and can be read normally.
The values provided are common starting points and should be measured against the reader's own samples.