What the data for this category looks like
Data for active pharmaceutical ingredient and pharmaceutical intermediate production process documents and exception reports comes from registered process documents, raw material batch ledgers, DCS production data, weighing system records, LIMS laboratory information management system, chromatography data system outputs, and QMS quality management system. Daily generated documents include batch production records, material balance sheets, reaction and purification step records, cleaning validation records, impurity profile reports, deviation and change documents. Updates trigger with each production batch and completed process step. Documents include fields such as batch number, reaction yield, impurity content, solvent residue amount, and process parameters. They present hierarchical characteristics aligned with multi-step synthesis or fermentation workflows.
What constraints these characteristics impose on deployment and upgrade
Data for this category is multi-source, heterogeneous, and updates frequently. Each production batch generates new documents and data. Deployment requires configuring parameters that support importing data from multiple systems. Hierarchical long documents need matching storage and index configurations to avoid overloading during single loading. Content such as impurity profile reports and change documents involves registered compliance-related data. Deployment requires configuring parameters for permission control and encrypted storage. Upgrade steps must ensure complete data migration. Version updates must not cause field mapping failures between LIMS and chromatography data systems, which would reduce the accuracy of exception report drafting.
How to set the configurations
| Configuration Item | Recommended Setting | Rationale for This Setting |
|---|---|---|
MONGODB_URI | Replica set connection string for local private deployment, e.g. mongodb://user:pass@192.168.1.50:27017/fastgpt?authSource=admin | Store hierarchical long documents such as batch production records and impurity profile reports, to meet the data storage needs that increase with production batches for this category |
REDIS_URL | 192.168.1.51:6379?password=secure_pass_123 | Cache frequently updated process step data and user sessions, to reduce database pressure during concurrent access across multiple departments |
STORAGE_S3_ENDPOINT | Intranet S3-compatible storage endpoint, e.g. http://192.168.1.52:9000 | Store large-volume raw data files exported from chromatography data systems, to meet compliance requirements for offline storage |
FILE_TOKEN_KEY | 32-bit randomly generated string | Generate temporary download tokens for process documents and exception reports, to control access permissions for sensitive data |
DB_MAX_LINK | 20–30 | Adapt to concurrent database connection requests from synthesis, quality, registration, and other departments, to avoid service interruptions caused by exhausted connections |
ROOT_KEY | Strong password string for internal operation and maintenance use only | Control access to the platform's highest permissions, to prevent unauthorized personnel from modifying configurations related to process data |
Three common mistakes
- Symptom: Updated images cannot be pulled after offline deployment, and the service version lags. Cause: The FastGPT v4.17.0 offline image package was not exported in advance, and no local image repository was configured for version upgrades.
- Symptom: Impurity profile data fields are empty after knowledge base migration, and exception reports cannot be generated. Cause: Original field mapping configuration exported from the LIMS system was not retained during migration, causing data to fail correct association with process document templates.
- Symptom: Frequent database connection timeouts occur during single-machine deployment, with error status code 503. Cause: The
DB_MAX_LINKconfiguration is too low, failing to adapt to concurrent process document query requests across multiple departments, leading to exhausted connection pools.
How to confirm configurations are correct
- Run the
docker pscommand. Confirm that MongoDB, Redis, S3 storage dependency containers, and the FastGPT main container are all running normally. - Upload a test impurity profile report file. Verify that the file can be stored normally and a temporary download link is generated. Confirm that the
FILE_TOKEN_KEYconfiguration takes effect. - Initiate process document query requests matching actual production concurrency levels. Verify that database connection counts do not reach the upper limit. Confirm that the
DB_MAX_LINKconfiguration adapts to business needs. - Log in to the platform using the
ROOT_KEY. Verify that the highest permission settings page is only accessible via this key. Confirm that the permission control configuration takes effect.
The values given are common starting points and should be measured against the reader's own samples.