What information is flushed when the chunk-size value is changed in the config dlp settings?

DLP document fingerprinting works by breaking content into fixed-size 'chunks' and hashing them to create a unique identifier. If the `chunk-size` is modified, the method of breaking down documents changes, rendering all previously generated fingerprints incompatible and requiring the database to be flushed so new fingerprints can be created.

The `chunk-size` setting defines how content is analyzed for fingerprinting, not which file types are supported by the DLP filters.

Archived files and messages are historical records of DLP events; changing the `chunk-size` does not affect already stored content or their archiving status.

File name patterns are used for identifying files based on their names, which is a separate mechanism from the internal content analysis and chunking used for document fingerprinting.