Data for this category
Data is generated across the full version release cycle. Sources include version tags in Git repositories, merged commit records, artifact packages produced by CI/CD workflows, and associated Issue milestones. Daily maintained documents include release notes, change logs, version tag documentation, and pre-release checklists. Data updates occur once per version release. Core fields include version number, change type classification, artifact hash value, and associated milestone number. The core output of this category is a publicly released set of merged changes, which has a clear distinction from single-change validation in code review.
Constraints for deployment and upgrade
Version data for this category is updated in batches per version cycle. Deployment must bind the corresponding version’s artifact package and metadata to avoid configuration misalignment caused by cross-version mixing. Multiple-source data validation requires configuring a unified metadata storage link during deployment, to ensure consistency between version tags, change records, and artifact packages. Preparations for change classification and destructive change descriptions require loading the corresponding version’s change documents during upgrade, for version matching verification during troubleshooting. The release artifact verification process requires configuring access permissions for artifact storage during deployment, to prevent inability to load static resources related to release notes after upgrade.
Configuration settings
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
MONGODB_URI | Supports single-node or replica set format. For single-node, use mongodb://<host>:27017/fastgpt | Adapts to database connection requirements for FastGPT v4.17.0, compatible with local deployment and cloud environment database configurations |
REDIS_URL | redis://<host>:6379 | Uses the default Redis port, adapts to FastGPT’s cache and session management logic |
DB_MAX_LINK | 100–200 connections | Balances database load and concurrent request processing capacity, avoids service interruption from exhausted connections |
STORAGE_S3_ENDPOINT | Fill in the access address provided by the object storage service provider | Used to access artifact packages and release notes static resources for version releases |
FILE_TOKEN_KEY | Randomly generated 32-bit string | Used to verify the legitimacy of file upload requests, prevents unauthorized file submissions |
Three common configuration errors
- Symptom: MongoDB connection failure after deployment, logs indicate missing replica set configuration. Cause: Failed to configure correct connection parameters per
MONGODB_URIrequirements, mistakenly forced replica set parameters for a single-node database, causing connection verification failure. - Symptom: Data processing is empty after uploading knowledge base files, the interface shows no processing tasks. Cause:
FILE_TOKEN_KEYis not configured, or the parameter value does not match the signature verification rules of the local storage service, causing file upload requests to be blocked. - Symptom: After upgrading the version, release notes for some historical versions cannot be loaded, and the frontend returns a 404 error. Cause: The
FE_DOMAINconfiguration was not updated, causing static resource paths to point to the old deployment address, which cannot match the new version’s file storage location.
How to verify correct configuration
- A database connection test command is executed to verify connectivity for
MONGODB_URIandREDIS_URL, confirming there are no connection timeout or permission denied errors. - A test file under 10MB is uploaded, and the backend data processing queue is checked for corresponding tasks, confirming the file upload and processing link is working normally.
- The configured
FE_DOMAINaddress is accessed, verifying that the frontend page loads normally, and the release notes module can display the corresponding version’s document content. - System runtime logs are reviewed, confirming that the connection count for
DB_MAX_LINKdoes not trigger alarm thresholds, and the database connection pool status is stable.
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.