This guide is compatible with FastGPT v4.17.0.
What this category of data looks like
Data for the basic chemical continuous production category primarily comes from DCS/SIS systems, online analyzers, LIMS laboratory management systems, equipment condition monitoring modules, energy metering terminals, and raw material tank data interfaces. Daily documents include operating procedures, process indicator cards, DCS trend charts, laboratory reports, interlock bypass records, special work permits, and shift logs. Data update frequencies fall into three categories: real-time, per shift, and daily. Document fields include unit-attached process parameters such as temperature in ℃ and pressure in MPa, interlock numbers, work approval numbers, and other unique identifiers.
What constraints these characteristics impose during deployment and upgrade
Multiple data sources with significantly different update rhythms require differentiated configuration for multi-source data synchronization during deployment, to avoid index conflicts across data of different frequencies. Document types cover structured parameter tables and unstructured trend charts, so compatible multi-format parsing rules must be configured to ensure accurate field extraction for process indicator cards and text recognition for exception reports. Data related to critical interlocks has high real-time requirements, so hot update interfaces must be reserved during the upgrade phase to avoid impacts on production stability from downtime maintenance. The volume of documents for large-scale continuous production accumulates with the device operation cycle, so storage expansion paths must be planned during deployment to meet long-term data archiving needs.
How to set the configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
MONGODB_URI | Configure as a dedicated business database address, set connection pool size to 20–30 | Document and exception report data volume for basic chemical continuous production accumulates with operation cycles, requiring sufficient connection pool to support concurrent read and write operations |
PG_URL | Configure as the database address for storing vector data and structured process parameters, set connection timeout to 600 seconds | Stable PostgreSQL connections are required for structured process indicators and interlock data; 600-second timeout supports bulk historical data import |
REDIS_URL | Configure as a dedicated cache address, set maximum connections to 15 | Cache real-time alerts and shift report synchronization status; 15 connections meet concurrent query requirements across multiple workstations |
STORAGE_S3_ENDPOINT | Configure as the enterprise internal object storage endpoint | Store large-volume unstructured documents such as DCS trend charts and laboratory reports, adapting to the enterprise’s existing storage architecture |
DB_MAX_LINK | Set to 50–70 | Meet concurrent connection requirements for multi-source data synchronization and exception report retrieval, avoiding database connection exhaustion |
ROOT_KEY | Generate a key in accordance with internal enterprise permission specifications | Ensure the security of system background management permissions, complying with permission control requirements in industrial scenarios |
Three common mistakes
- Symptom: After uploading process indicator cards and DCS trend charts, the corresponding document content cannot be returned during retrieval. Cause: Parsing rules compatible with multi-format documents are not configured, leading to failure to correctly index unstructured documents.
- Symptom: After configuring
PG_URL, the system displays a maximum of 30 createable knowledge bases. Cause: The vector storage sharding parameters of the PostgreSQL database are not adjusted; the default configuration limits the number of createable knowledge bases. - Symptom: Restart fails after Docker deployment, with logs showing database connection timeout. Cause:
REDIS_URLis not configured as a dedicated cache address, and shared cache with other services leads to exhaustion of connection resources.
How to confirm correct configuration
- Run the database connection test script to verify the connectivity of
MONGODB_URIandPG_URL, confirming no timeout errors. - Upload a typical shift log and process indicator card, perform keyword retrieval, and confirm that the corresponding documents are correctly recalled.
- Check the system operation logs to confirm that the number of cache connections for
REDIS_URLdoes not exceed the configured upper limit, with no connection exception alerts. - Attempt to modify system background configuration items to confirm that
ROOT_KEYverification takes effect, and the permission control logic meets expectations.
The values provided here are common starting points, and should be evaluated against the reader’s own samples.