What the data for this category looks like
The data for this category primarily comes from Git repositories, CI pipelines, static scanning reports, and defect databases. Update frequency is daily or per commit. Regularly generated documents include merge requests, code review comments, commit records, and branch strategy descriptions. A single merge request entry includes a list of changed files, assigned reviewers, commit hashes, and CI execution status fields. Code review comments include commenter identifiers, comment timestamps, and review conclusion tags.
What constraints these characteristics impose on the conversation logs and auditing workflow
Since data sources cover code repositories, pipeline tools, and defect management systems, the conversation logs and auditing workflow must correlate multi-source heterogeneous data to ensure full change traceability during troubleshooting. The high-frequency update feature requires audit logs to support real-time synchronization, avoiding missing latest review or merge records during incident troubleshooting due to delays. Differences in fields across document types require prioritizing correlation of merge request change scopes and core review conclusions when assembling context, to reduce interference from irrelevant information. Additionally, the serious consequences of defects entering the main branch require audit logs to retain full review, merge, and rollback workflow data, without truncating critical steps.
How to configure settings
| Configuration Item | Recommended Approach | Rationale |
|---|---|---|
chatId | Bind to merge request ID, retain for 180–365 days | Audit cycles for merge requests typically cover their full lifecycle. Binding chatId to merge request IDs allows quick retrieval of corresponding conversation logs via Git repository merge request IDs |
responseChatItemId | Generate a unique identifier for each review comment, increment in commit order | Allows precise tracing of conversation context for individual review comments, avoiding confusion across multiple review rounds |
detail | Store full CI scan results, review comments, and merge status, do not truncate content over 2000 characters | Complete details support change impact analysis during incident troubleshooting, preventing loss of critical information due to truncation |
| Audit log sync trigger rules | Trigger sync per commit, delay no more than 60 seconds | Matches the high-frequency update rhythm of this category, ensuring latest review and merge records are available during incident troubleshooting |
| Conversation log retrieval limit | Return the first 20 associated records | Reduces interference from excessive irrelevant data, while covering most review rounds for merge requests |
Three common configuration mistakes
- Symptom: The conversation details page cannot display internal review communication records, and returns empty content. Cause:
chatIdis not bound to the merge request ID, so retrieval cannot associate the corresponding data source. - Symptom: No conversation history records exist in the MongoDB collection. Cause: Audit log persistence configuration is not enabled, or retention period is set to 0 days, causing automatic data cleanup.
- Symptom: Index creation prompts that content exceeds length limits. Cause: No truncation rule is configured for the
detailfield, and excessively long merge request changed file lists or review comments are stored directly.
How to verify correct configuration
- Access the FastGPT v4.17.0 admin backend, navigate to the application configuration page for the corresponding code review and merge management workflow, and verify that
chatIdis bound to the merge request ID field. - Initiate a simulated code commit and trigger a merge request, view the audit log list, and confirm that review comment records corresponding to
responseChatItemIdcan be retrieved. - Access the MongoDB database, check for the presence of conversation log data containing the
detailfield in the corresponding collection, and confirm that field content is not excessively truncated. - On the merge request details page, click the associated conversation log link, and confirm that the full review communication record page loads correctly.
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.