What the data for this category looks like
Data for this category primarily comes from customer ITSM systems, operations monitoring platforms, internal asset inventories, and on-duty records. Daily generated documents include on-site daily reports, service desk work orders, inspection reports, and SLA monthly reports. Document fields cover work order number, processing duration, inspection item status, SLA compliance rate, and other items. Operations staff typically update on-duty and work order records daily, and compile inspection and SLA data monthly. Data formats are usually a mix of structured tables and unstructured documents. Data volume changes dynamically with the scale of on-site services.
Constraints imposed by these characteristics during deployment and upgrade
Data sources are scattered and involve the customer’s existing ITSM and monitoring platforms. During deployment, configure support for multi-source data access, and avoid conflicts with the customer’s existing system permission rules. Document structures are diverse and fields are inconsistent. During upgrade, support compatibility with different formats of on-site daily reports and work order data, and ensure data parsing logic can be adjusted flexibly. Differences in update frequencies require deployment to support flexible scheduled synchronization task configuration, to adapt to daily and monthly update rhythms. Additionally, on-site services involve customer-site deployment. The upgrade phase must support non-stop or low-downtime rolling updates, to avoid disrupting normal customer operations on-duty workflows. Simplify the configuration process to lower the operation threshold for on-site engineers.
Configuration guidelines
This configuration is based on FastGPT v4.17.0.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
MONGODB_URI | mongodb://username:password@internal-mongo:27017/fastgpt?authSource=admin | Adapt to the scattered work order and on-duty record storage needs of on-site services, support association of multiple document types |
REDIS_URL | redis://:redis-password@internal-redis:6379/0 | Used to cache frequently queried work order status and on-duty information, reducing database load |
DB_MAX_LINK | 20–30 | Match the concurrent access volume of daily work orders and inspection data for on-site services, avoid connection exhaustion |
FE_DOMAIN | https://fastgpt.internal.customer.com | Ensure on-site engineers can normally access the Q&A and work order management interfaces in the intranet environment |
STORAGE_S3_ENDPOINT | http://internal-s3.example.com:9000 | Used to store document attachments such as on-site daily reports and inspection reports, adapt to the customer’s existing storage specifications |
ROOT_KEY | Randomly generated 32-bit string | Used to configure interface call permissions for custom plugins, prevent unauthorized access |
Three common configuration errors
- Phenomenon: Q&A results do not include the configured knowledge base content, only return generic responses. Cause: The vector model and knowledge base associated storage parameters are not configured correctly, resulting in failed document index generation.
- Phenomenon: After adding a custom plugin to the workflow, no input or output parameters are displayed, and no error logs are generated. Cause: The correct
ROOT_KEYpermission is not configured, resulting in insufficient plugin interface call permissions, and parameters cannot be loaded normally. - Phenomenon: Continuous errors are returned when calling the embedding model. Cause: The self-built embedding model is not connected to the specified API channel, or the
STORAGE_S3_ENDPOINTis not configured to adapt to the model’s storage path.
How to verify successful configuration
- Access the intranet address corresponding to
FE_DOMAIN, verify that the interface loads normally, with no resource loading failure or permission error prompts. - Upload a standard-format on-site daily report document, verify that the document can be correctly parsed and indexed to the target knowledge base.
- Trigger a work order query workflow, verify that the custom plugin can load normally and display preset input and output parameters.
- Call the embedding model interface, verify that the returned result format meets expectations, with no connection timeout or permission exceptions.
The values provided are common starting points and should be measured 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 2026-10-19.