Unable to parse date in grok and filters

Elastic Stack Logstash
#1

I have the following message

2021-01-04T00:04:00.8033345+00:00 0HM5E5E28610F:00000001 [INF] MyMessage

What I'm trying to do is to extract each of those fields into a separate column for my index.

This is what I have been using to parse the message in my pipeline

    filter {
    mutate {
      strip => "message"
    }
    grok {
      match => { "message" => "%{DATA:logdate} %{DATA:thread_id} \[%{LOGLEVEL:log.level}\] %{GREEDYDATA:message}" }
    }
    date {
      match => [ "logdate", "yyyy-MM-dd HH:mm:ss.SSSSSSSZZ" ]
      target => "@timestamp"
    }
}

At the end, I'm getting parse errors and my message ends up in Elasticsearch by sending the whole message as a column in my index. The idea is to break the 4 data I have in each line into different columns. Not only GrokParseError I receive, but also issues while converting the date.

Any idea on how to solve it?

#2

I think it should be worked with grok.

%{TIMESTAMP_ISO8601:action} %{NOTSPACE:thread_id} \[%{WORD:log.level}\] %{GREEDYDATA:message}
1 Like
#3

Thank you very much, it could solve the parsing issue now.
Although, I still see 3 issues:

image
image716×623 29.3 KB

  1. The @logdate field shouldn't exist, as I want to replace the log entry time from the @timestamp, that means, my timestamp should be pointing to January, not the collected date;

  2. The message should contain only the last field in my log entry, expected would be only "msg"

  3. I see that the additional fields I added are giving me an exclamation sign that the field is not cached

Any ideas how to solve the above?

Thanks!

#4

(3). Go to Management, index pattern and refresh field. That is quick way. But you can index template (it look like database schema on mysql/sql).
(2). I think no need replace "message" field. You can remove message and create new field msg
(1). You can use date filter plugin https://www.elastic.co/guide/en/logstash/current/plugins-filters-date.html#plugins-filters-date-match

filter {
    grok {
        match => { "message" => "%{TIMESTAMP_ISO8601:action} %{NOTSPACE:thread_id} \[%{WORD:log.level}\] %{GREEDYDATA:msg}" }
    }

    mutate {
        remove_field => [ "message" ]
    }
}
1 Like
#5

Thanks, that did the trick!

Best Regards.