HTML and XML files encoded as UTF-16 LE may be deleted by the DLP Endpoint Agent during detection.
Affects the following configurations:
The issue is caused by an optimization on Windows for handling UTF-16 LE (Little Endian) encoded files during content extraction.
Set the Detection.MARKUP_AS_TEXT Advanced Agent Setting back to the default value of OFF.
A fix for this issue is planned for the 15.8 MP3 release of the DLP Endpoint Agent.