What the data for this category looks like
Data for this category comes from the Chinese Center for Disease Control and Prevention Information System, hospital HIS systems, infectious disease reporting systems, and laboratory testing systems. The update frequency is daily. Daily processed documents include infectious disease reporting cards, epidemic analysis reports, case correction records, clustered epidemic reports, and epidemiological investigation reports. Fields include basic case information, diagnosis results, reporting time limits, correction notes, test results, and similar items. All documents must strictly follow legal formats, with no custom extended fields.
What constraints do these characteristics impose on deployment and upgrades
This category uses data from multiple heterogeneous systems. During deployment, multiple data source connection configurations must be set up to ensure synchronization tasks can stably pull documents from different systems. The daily update frequency requires scheduled synchronization tasks to be configured during deployment, and sufficient concurrent processing resources must be reserved. Documents must strictly follow legal formats, so the vector database field storage structure must be matched during deployment to avoid data format incompatibility. Public health data has severe consequences for errors, so high-availability mode must be configured during deployment to ensure uninterrupted service. Data encryption and permission control rules must also be configured to protect data security.
How to set configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
PG_URL | jdbc:postgresql://<host>:5432/fastgpt?currentSchema=public | Stores structured data such as infectious disease reporting cards and case correction records, adapts to daily high-frequency write requirements |
REDIS_URL | redis://:<password>@<host>:6379/0 | Caches daily updated epidemic summary data to reduce repeated request pressure on heterogeneous systems |
MONGODB_URI | mongodb://<user>:<pass>@<host>:27017/fastgpt?authSource=admin | Stores unstructured documents such as epidemiological investigation reports and epidemic analysis reports, supports large field storage |
DB_MAX_LINK | 10–15 | Handles concurrent requests from daily multi-system synchronization to avoid database connection exhaustion |
FILE_TOKEN_KEY | Randomly generated 32-bit string | Encrypts access tokens for case-related files to protect data access permissions |
ROOT_KEY | Complex random string, rotated every quarter | Verifies the highest permissions of the management backend to prevent unauthorized access |
Three common mistakes
- Phenomenon: After upgrading FastGPT v4.17.0, the model thinking output switch is enabled, but the <think></think> tags still appear in the main text. Cause: The front-end configuration was not resynced after the upgrade, or
FE_DOMAINwas not correctly configured during deployment, causing abnormal front-end rendering logic. - Phenomenon: When deploying for the first time, after selecting a storage version, existing data cannot be easily migrated. Cause: No cross-storage engine data migration script was planned in advance, or the selected storage engine does not support batch export of structured data.
- Phenomenon: When passing the userId parameter via URL, the system cannot obtain the corresponding user identifier. Cause: The custom user ID pass-through function was not enabled in the deployment configuration, or the interface parameter name change was not synchronized to the business logic.
How to confirm the configuration is complete
- Log in to the FastGPT management backend, enter the database monitoring page, check the connection status of the
PG_URLandMONGODB_URIconfigurations, and confirm there are no abnormal error reports. - Manually trigger a daily data synchronization task, check whether infectious disease reporting cards, epidemic analysis reports and other documents can be pulled from the specified data sources, and whether the data format matches the preset structure.
- Call the test interface, append the userId parameter to the URL, and check whether the system log can correctly obtain and associate the operation records of that user.
- Enable the model thinking output switch, generate a test reply, and check whether the front-end page only displays the thinking process in the pop-up window, with no remaining <think> tags in the main text.
The values provided here are common starting points and should be evaluated against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.