What the data for this category looks like
Data for this category comes from the national medical security information platform, hospital HIS and medical insurance interfaces, medical insurance catalog and coding library, and cross-provincial medical treatment settlement platform. Daily processed documents include medical insurance settlement lists, expense details, medical record homepages, prescriptions, rejection payment lists, and settlement reconciliation statements. Data updates occur daily. Core fields cover diagnosis and treatment item and drug consumable codes, expense amount, settlement status, insured person identification, and more. Coding fields are fixed-length digits. Amount fields use yuan as the unit. Time fields use standard timestamp format.
What constraints these characteristics impose on deployment and upgrade
Multiple data sources and daily incremental updates require configuring multi-source data synchronization links during deployment. This avoids excessive computing resource usage caused by full data pulls. Document types include structured and semi-structured files. Different parsing rules must be adapted. The upgrade process must support newly added medical insurance settlement list formats. The core audit link relies on code matching and consistency checks. This imposes clear requirements for vector recall accuracy. Sufficient computing resources must be reserved during deployment. Data related to medical insurance funds is involved. Deployment must comply with data security specifications. The upgrade process must use gray release to avoid impacting daily settlement services.
How to set configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
PG_URL | postgresql://user:pass@host:5432/fastgpt?pool_max_size=20 | The average daily incremental volume of medical insurance data is large. Setting a reasonable connection pool size prevents database connection exhaustion |
REDIS_URL | redis://:pass@host:6379/0?maxmemory-policy=allkeys-lru | Used to cache medical insurance code matching results. The LRU policy prevents memory overflow |
STORAGE_S3_ENDPOINT | http://minio-host:9000 | Adapts to the bulk storage needs of medical insurance documents, compatible with S3 protocol local storage services |
DB_MAX_LINK | 20–30 connections | Matches the average daily data processing volume, prevents service interruptions caused by exhausted database connections |
FILE_TOKEN_KEY | Custom random 32-character string | Secures access to the file upload interface, prevents unauthorized calls |
Three common mistakes
- Symptom: After uploading documents such as medical insurance settlement lists and prescriptions, the file parsing node has no activity, and the interface shows the file as unparsed. Cause:
STORAGE_S3_ENDPOINTis not configured correctly, resulting in an invalid file storage path, and the parsing node cannot read the target file. - Symptom: The PostgreSQL container restarts continuously after Docker deployment. Checking the container logs displays the error "could not accept connection: too many clients already". Cause:
DB_MAX_LINKis set too small, orPG_URLdoes not have reasonable connection pool parameters configured, resulting in exhausted database connections. - Symptom: Dependency version conflict errors occur when building the FastGPT image, such as "npm ERR! could not resolve dependency". Cause: The official dependency list for FastGPT v4.17.0 is not used for installation, resulting in version mismatches.
How to confirm the configuration is correct
- Run
docker logs fastgpt-pgto check that the PostgreSQL container has no errors related to "too many clients already", confirming that thePG_URLandDB_MAX_LINKconfigurations take effect. - Upload a simplified version of the medical insurance expense detail test file, check that the application logs show normal extraction of core fields such as diagnosis and treatment item codes and expense amounts, confirming that
FILE_TOKEN_KEYand storage configurations are correct. - Access the party building learning Q&A application management page at
FE_DOMAIN, verify that the file upload function can be triggered normally and complete parsing, confirming that theFE_DOMAINconfiguration is correct. - Check Redis monitoring metrics, confirm that the cache hit rate meets business expectations, verifying that the
REDIS_URLconfiguration is valid.
The values provided 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.