@timestamp cannot be deleted - overwriting does not work

Elastic Stack Logstash
1

Hi,
I know, that the field @timestamp can not be deleted. So I want to override @timestamp with the timestamp of logs. I've created a new field. The timestamp in logfile looks like "13.07.2017 07:11:10 ..." my new field LOGTIMESTAMP "13.07.2017 07:11:10", but @timestamp "17.07.2017 11:29:24".
I parse the logfile in this way ->
grok {
match => [ "message", "%{DATESTAMP:LOGTIMESTAMP}" ]
}
date {
match => ["LOGTIMESTAMP", "dd.MMM.YYYY hh:mm:ss"]
target => "@timestamp"
locale => "de"
remove_field => [ "LOGTIMESTAMP" ]
}
What am I doing wrong? Thanks for your help.

2

If the date filter fails the log contains clues. I can immediately see that MMM in your date pattern should be MM and, unless you use a 12-hour clock with no am/pm marker, hh should be HH.

3

Thanks for your answer. I've made the adjustments as suggested, without success. I can not get any error messages from the logfiles. Kibana shows me an _dateparsefailure, otherwise no further error-informations ...

4

Works fine for me with both Logstash 2.3 and 5.4. Please copy/paste an example event from Kibana. Use the JSON tab in the Discover view so we get the raw JSON representation.

$ cat test.config
input { stdin { } }
output { stdout { codec => rubydebug } }
filter {
  date {
    match => ["message", "dd.MM.YYYY HH:mm:ss"]
    locale => "de"
  }
}
$ echo '13.07.2017 07:11:10' | logstash -f test.config
Settings: Default pipeline workers: 8
Pipeline main started
{
       "message" => "13.07.2017 07:11:10",
      "@version" => "1",
    "@timestamp" => "2017-07-13T05:11:10.000Z",
          "host" => "bertie"
}
Pipeline main has been shutdown
stopping pipeline {:id=>"main"}
5

If i reproduce your example, I get the same (correct) result. But only for manual execution.

Here my JSON-Output:

{
"_index": "filebeat-2017.07.18",
"_type": "test",
"_id": "AV1UtBnpzzBeDCd4SQD4",
"_score": null,
"_source": {
"HOST": "test.lan",
"PID": "20896",
"message": "13.07.2017 00:51:10 [20896] [INFO] [main::() line(63)] staging dirs:",
"tags": [
"_dateparsefailure"
],
"@timestamp": "2017-07-18T07:57:34.263Z",
"LOGMESSAGE": "staging dirs:",
"PRODUCT": "test",
"beat": {},
"SOURCE": "main::() line(63)",
"INPUT_TYPE": "log",
"FILE": "/tmp/test.log",
"LOGTIMESTAMP": "13.07.2017 00:51:10",
"TYPE": "test"
},
"fields": {
"@timestamp": [
1500364654263
]
},
"sort": [
1500364654263
]
}

6

I can't reproduce.

$ cat data
13.07.2017 00:51:10 [20896] [INFO] [main::() line(63)] staging dirs:
$ cat test.config 
input { stdin { } }
output { stdout { codec => rubydebug } }
filter {
  grok {
    match => ["message", "%{DATESTAMP:LOGTIMESTAMP}"]
  }
  date {
    match => ["LOGTIMESTAMP", "dd.MM.YYYY HH:mm:ss"]
    locale => "de"
  }
}
$ logstash -f test.config < data 
Sending Logstash's logs to /home/magnus/logstash/logstash-5.4.1/logs which is now configured via log4j2.properties
[2017-07-18T15:56:07,689][INFO ][logstash.pipeline        ] Starting pipeline {"id"=>"main", "pipeline.workers"=>8, "pipeline.batch.size"=>125, "pipeline.batch.delay"=>5, "pipeline.max_inflight"=>1000}
[2017-07-18T15:56:07,715][INFO ][logstash.pipeline        ] Pipeline main started
[2017-07-18T15:56:07,765][INFO ][logstash.agent           ] Successfully started Logstash API endpoint {:port=>9600}
{
      "@timestamp" => 2017-07-12T22:51:10.000Z,
        "@version" => "1",
            "host" => "bertie",
         "message" => "13.07.2017 00:51:10 [20896] [INFO] [main::() line(63)] staging dirs:",
    "LOGTIMESTAMP" => "13.07.2017 00:51:10"
}
[2017-07-18T15:56:10,733][WARN ][logstash.agent           ] stopping pipeline {:id=>"main"}

I find it very hard to believe that the date filter is silent about why it's failing.