What data for this category looks like
Data for this category primarily comes from CMDB platforms, cloud resource inventories, automated discovery tools, and procurement ledgers. Update frequency is daily or weekly. Commonly processed documents include configuration item lists, service component relationship topologies, change records, and asset inventory tables. Document fields include configuration item unique identifiers, asset models, dependency relationships, change times, and responsible persons. Some documents contain structured tables and hierarchical topologies.
How these characteristics impose constraints on document parsing and chunking
Documents in this category include structured tables, hierarchical topologies, and many repeated fields. The parsing workflow must support multi-format content, and avoid splitting cross-page topologies or associated tables. Daily or weekly update cycles require efficient parsing, and avoid long processing times. Configuration items and dependency relationships have strong binding properties. Chunking must retain contextual associations, to prevent loss of component dependency links after splitting. Document formats vary widely across sources. The system must support multiple content forms, including plain text, tables, and nested structures.
Configuration settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Aligns with the field length of configuration item lists and asset inventory tables, and avoids splitting critical dependency relationship descriptions |
chunkSplitter | Split by paragraph + table boundaries | Retains the association between configuration items and corresponding asset information, and prevents incorrect splitting of table content |
customPdfParse | Enabled | Adapts to PDF topology diagrams and structured tables exported from CMDB, and improves parsing completeness |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Supports bulk exported annual asset inventory tables and combined change records |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Meets the parsing time requirements for large topology documents, and avoids timeout interruptions |
chunkSettingMode | Manual configuration | Adjust chunking rules for documents from different sources (such as CMDB lists, cloud resource inventories) to adapt to exclusive data structures |
Three common configuration mistakes
- Uploading a PDF exported from CMDB results in empty parsed content or only scattered text. The cause is that the
customPdfParseconfiguration is not enabled, so the topology structure of the dedicated export format cannot be recognized. - Parsed tables from asset inventory sheets have misaligned cells or jumbled content. The cause is that
chunkSplitteris not configured to split by table boundaries, leading to unordered splicing of cross-page table content. - Batch uploading multiple change records results in a timeout error. The cause is that
PARSE_FILE_TIMEOUT_SECONDSis set lower than the actual processing time, and does not adapt to the parsing requirements of large documents.
How to confirm correct configuration
- Upload a standard CMDB configuration item list PDF, and check if the parsed text retains the association between configuration item IDs and corresponding asset models.
- Upload an asset inventory sheet containing cross-page tables, and verify that the parsed table content has no misalignment or missing content.
- Upload an 800 MB combined change record set, and confirm that the parsing process does not trigger a timeout error.
- In FastGPT v4.17.0, adjust
chunkSizeto the 800–1200 character range, compare the contextual integrity of the chunks, and confirm that it meets the field length requirements of business data.
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.