What this category of data looks like
Core data for this category comes from incident timelines, system logs, monitoring metrics, code commit records, and test case libraries. Daily produced documents include RCA reports, incident tickets, review minutes, and rectification tracking sheets. RCA reports contain structured fields such as incident occurrence time, associated test case ID, defect number, root cause analysis conclusion, and rectification measures. Data updates are triggered by major incidents, supplemented by monthly fixed review updates. The overall data structure focuses on incident correlation and rectification closure.
Constraints on citation sources and traceability
Mixed multi-source data requires the traceability link to accurately match unique identifiers of different data sources. This avoids information confusion across incidents and test cases. Structured document fields require locking specific field ranges during citation. This ensures accurate correlation between associated test cases and defect information. Non-fixed update frequency requires traceability to bind document versions from corresponding time nodes. This prevents citation of expired or irrelevant rectification records. The core goal of this category is root cause location and prevention. The accuracy of the traceability link directly affects rectification effectiveness. Redundant manual verification steps should be reduced.
Configuration guidelines
This configuration applies to FastGPT v4.17.0 and later versions. Ensure that AIPROXY_API_ENDPOINT (root address of the AI Proxy service, e.g., http://aiproxy:3000; do not append /v1 or /chat/completions, do not configure as a single-model address such as VLLM/Ollama/Xinference) and AIPROXY_API_TOKEN (AI Proxy administrator token, consistent with the AI Proxy server configuration and not empty; do not use application API keys, model vendor keys, application-issued tokens, or request-based token count estimates) have been correctly configured in the App/Admin backend. Incorrect or missing configuration will cause startup verification failure. This configuration is the foundation for normal operation of the citation traceability function, and cannot be replaced by workflow global variables, application API request headers, or ordinary model management pages. Upstream model channel addresses and vendor credentials must be handled separately via model channel configuration.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
| Maximum citation token count | 8000–12000 token | Matches the total token scale of RCA reports and associated logs for this category, avoids truncation of critical traceability information |
| Response data field scope | Full set including associated test case ID and defect number | Ensures traceability can accurately correlate to specific test cases and defect entries |
| Detailed context switch | Enabled | Supports retrieval of multi-source detailed data such as incident timelines and code commit records, supports root cause hypothesis verification |
| Single recall limit | Top 5 entries | Matches the typical number of associated documents reviewed monthly for this category, avoids redundant information interfering with analysis |
| Citation source display switch | Configured as needed | Adapts to display habits of different teams, complies with document output specifications for incident management |
Three common configuration errors
- Forcing citation sources to display in conversation response results, with no option to hide as needed. Troubleshooting steps: Check the configuration value of the citation source display switch, check whether non-category-specific document citation rules are bound.
- Traceability results correlating to incorrect test cases or defect entries. Troubleshooting steps: Check whether the response data field scope configuration includes the full set of associated test case ID and defect number.
- Cited incident log content being truncated, with no access to complete root cause analysis basis. Troubleshooting steps: Check whether the maximum citation token count configuration matches the total token scale of RCA reports and associated logs.
How to confirm correct configuration
- Initiate a query containing a known associated test case ID, check whether the returned results include traceability information for the corresponding test case.
- View the citation source module of the returned results, confirm that the configured associated field range is included.
- Test a query for a long RCA report, confirm that content is not unnecessarily truncated.
- Adjust the value of the citation source display switch, verify that the display status in the results updates synchronously.
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.