This document applies to the deployment and upgrade of the SLA and service reports module for FastGPT v4.17.0.
What the Data for This Category Looks Like
Data for this category comes from monitoring systems, ticketing systems, contract texts, and customer communication records. Corresponding documents include SLA clauses, monthly service reports, availability statistics reports, and customer retrospective minutes. Data is updated on a monthly or quarterly cycle. Core fields cover service availability metrics, root cause analysis results of abnormal events, contractually agreed SLA thresholds, customer feedback records, and more. Configurations must strictly align with contract terminology and actual service performance. No additional derived statistical fields are allowed.
Constraints Imposed by These Characteristics on Deployment and Upgrade
This category’s data relies on multi-source system access and periodic aggregation. Deployment must therefore support configuration of multiple storage media, while reserving flexible parameters for terminology adjustments. Monthly or quarterly update cycles do not require real-time high concurrency. However, storage node stability must be ensured to avoid data loss during periodic aggregation. Fields must strictly match contract agreements. Deployment configuration must support custom metric mapping to avoid terminology deviations caused by hard coding. Service report generation relies on a complete data source link. Any abnormal link will trigger compensation disputes. An abnormal alarm and retry mechanism must therefore be configured.
Configuration Recommendations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
MONGODB_URI | mongodb://username:password@host:27017/sla_db?authSource=admin | This category requires storing long-term SLA metrics and service report data, which needs persistence and support for complex queries. The default port is 27017. |
REDIS_URL | redis://:your_password@host:6379/0 | Used to cache monthly aggregated SLA metrics to reduce repeated calculation overhead. The default database index is database 0. |
DB_MAX_LINK | 20–30 connections | The data update frequency of this category is monthly or quarterly, so high-concurrency connections are not required. Excessive connections will occupy server resources. Adjust based on actual testing. |
STORAGE_S3_ENDPOINT | Fill in the endpoint provided by the cloud storage service provider of the deployment environment | Used to store raw service report data and archived files. Must adapt to the storage solution of the deployment environment. |
FE_DOMAIN | Fill in the intranet access domain name for intranet deployment, and the public network access domain name for public network deployment | Used to generate access links for service reports. Must match the customer’s actual access path. |
Three Common Mistakes
- Phenomenon: An
EMFILE: too many open fileserror occurs when executingdocker-compose up. Cause: TheDB_MAX_LINKconfiguration value is too high, exceeding the maximum number of file handles allowed by the system, resulting in exhausted database connections. - Phenomenon: The SLA metric field is empty after the service report is generated. Cause: The read/write permissions of
MONGODB_URIare not correctly configured, making it impossible to read availability data synchronized from the monitoring system, or the data synchronization link is interrupted. - Phenomenon: No vector generation results are displayed when importing text during intranet deployment. Cause:
FILE_TOKEN_KEYis not configured to generate valid file access tokens, causing the vector model to fail to read the imported text files.
How to Verify Successful Configuration
- Execute
docker psto confirm that all dependent service containers are in the Up state, with no abnormal restart records. - Call the SLA metric calculation interface to check whether the returned fields include the core metrics agreed in the contract, with no missing items.
- Generate a monthly service report to confirm that the data source annotations in the report match the actually connected monitoring and ticketing systems.
- Trigger an abnormal event root cause analysis task to confirm that results can be written normally to the specified storage medium, with no error logs.
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.