Extracting values from log body

Hi there,

Suppose I am getting a log message through logstash, with some custom format of the form FIELD1=VALUE1 etc. I am still quite a bit in the dark here, but I've been playing a little with visualizations. What I would like to do is obtain a unique count of, say, values of FIELD1. I think this is called an aggregation what I'm trying to do.

My question is: do I need to grok these fields at the logstash stage, and put all values in separate fields? Or can the field be parsed at search-time, Splunk-style? Or is there another recommended solution? I'm looking for "the ELK way" here.

Thanks for your thoughts.


Hi Dan,

I've moved your question to the Logstash forum as it seems to be the more appropriate place for this question.


Hi Daniel, and thanks. I honestly wouldn't know, hopefully it will get answered there.

If I understand your use case properly, I would try the kv filter, it seems the most suited to answer your needs here.


  • purbon

Hi Pere, and thanks for the answer. It doesn't really address my question though.
I am using kv, of course. The question is, do I need it?

The fields need to be extracted at index time in order for them to be available for aggregations in Kibana.

Thanks Christian! This answers my question.