How filebeat process memory mapped log files

I have found really strange behaviour of filebeat. Filebeat stop harvesting memory mapped files. I'm using log files which has constant size. To write data to logs i'm writing just string to memory and kernel flush it upon kernel finish write dirty pages. I found observation that filebeat can't collect all data from such kind of files. Filebeat skips a lot of lines in general it collect about 70% of content. Any ideas how it can be improved?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.