Issue parsing db2diag log using grok filter

Hi all,

I am new in ELK solution and currently I am working on Logstash -> Elasticsearch -> Kibana.
I need to parse db2diag.log. I am using grok filter for this. The problem is when I use grok debugger website everything work fine but when I put same filter on logstash _grokparsefailure happen. My log files looks like this:


PID : 5505958 TID : 258 PROC : db2ckpwd 0
INSTANCE: db2inst1 NODE : 000
EDUID : 258 EDUNAME: db2wdog 0 [db2inst1]
FUNCTION: DB2 UDB, oper system services, sqloSpawnAndWaitForPasswordCheckExe, probe:130
MESSAGE : ZRC=0x800F006A=-2146500502=SQLO_BAD_USER "Bad User"
DIA8117C Error with userid "".


And my grok filter:

(?m)%{GREEDYDATA:DATE} .*?:%{LOGLEVEL:loglevel}.*?:%{GREEDYDATA:pid}.*?:%{GREEDYDATA:tid}.*?:%{GREEDYDATA:proc}\n.*?:%{GREEDYDATA:instance}.*?:%{GREEDYDATA:node}\n.*?:%{GREEDYDATA:hostname}\n.*?:%{GREEDYDATA:eduid}.*?:%{GREEDYDATA:eduname}\n.*?:%{GREEDYDATA:function}.*?:%{GREEDYDATA:probe}\n.*?:%{GREEDYDATA:msg}

Is this correct. Do I have to use other filter then grok for this kind of parsing.


A pattern with twelve DATA and twelve GREEDYDATA is going to be ridiculously expensive if it fails to match. It may well exceed the 30 second timeout. I would suggest you match multiple patterns to pick out parts of the message

    grok {
        break_on_match => false
        match => {
            "message" => [
                "PID : %{NOTSPACE:pid}",
                "TID : %{NOTSPACE:tid}",
                "INSTANCE: %{NOTSPACE:instance}",
                "NODE : %{NOTSPACE:node}",
                "MESSAGE : %{GREEDYDATA:msg}"


Thanks a lot for guidance and quick response. I will try and confirm.

Thanks Badger. The log are parsed perfectly.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.