I understand multiline option could be exercised at filebeat level, but my question is when we are tokenizing the entire message in 5 distinct tokens, shouldn't the complete message get displayed as %{[message][4]}" ?
Because %{[message][4]}" is long text with a number of lines, it results in it not showing completely. Also the lines are not the same in number and pattern every time.
thanks. So in Kibana it does not show [message][4] as a complete message.
It seems to break it down into different sub-messages. Also as [message][4] is non-standard so different responses are seen. In the case below, it seems that it has taken the message (which is well within the token) as new line and lost formatting.
I expect the entire message (regardless of complexity) to appear completely as [message][4]
This image shows that your message is completely different from the one you shared in your first post.
As you see the value for the message field is just "name": "displayPassword",, so there is nothing to split because the string ~|~ does not exist in your message, so you will not have [message][X].
It seems that your source message is a multiline message, you would need to configure filebeat to consume the multiline message as one single message.
Yes the content of message field was different the second time around. The message comes in various sizes and formats, because of which I thought I would be able to encapsulate the entire message within one delimiter i.e. ~|~
Also you are right that the message in stdout is a multiline message. But as this message does not conform to a set format, it is hard to configure multiline message pattern in filebeat.
Is there a "one size fits all" kind of multiline pattern that I could follow at the filebeat level?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.