Hello,
I've got log files that can contain plain text lines from a Java application and JSON lines.
To complicate matters the plain text lines can be multiline , for example a Java stack trace, so I'm using the multiline codec in the input configuration. This doesn't play nice with the JSON lines as they match the multiline codec's pattern.
I was wondering how I can handle this in logstash?
I've tried:
- conditionals in the input, but this doesn't work and isn't designed to work.
- multiple pipeline to pipeline communication with an first pipeline detecting
{
and distributing to a JSON or Plain text pipeline, but I cannot use the multiline codec on the second pipeline, obviously as event could appear out of order.
Anyone have any other ideas I could test out?