Logstash event ruby code

hsiuming · November 22, 2017, 9:22am

Hello

We now set our own field by ruby code in logstash.

There are three Binary fields.

We have two question about:

(1) The logstash setting code is:

event.set('[Binary1]', event.get('[event_data][Binary]')[0..7])

event.set('[Binary2]', event.get('[event_data][Binary]')[8..15])

event.set('[Binary3]', event.get('[event_data][Binary]')[16..23])

if the Binary1 is not equal to "00000000",how can we use "event.remove('[Binary1]')" to remove this event.set?

I use "if [Binary1]=="00000000" {event.remove('[Binary1]')}",but it can not work.

(2) If the Binary1 is not equal to "00000000",we want to sent the data to elasticsearch.

The field rows of Binary will only have Binary1 not three fileds in the same data on kibana page.

We want to split the log data by Binary fields.

If we have three Binary fields, we will see three data on kibana page.

Could we set the logstash configure to do this ?

magnusbaeck · November 22, 2017, 9:57am

I use "if [Binary1]=="00000000" {event.remove('[Binary1]')}",but it can not work.

Use event.get to get field values. Secondly, Ruby doesn't use curly braces in this way. Or do this operation outside the ruby filter. Conditionally removing a field depending on its value is doable without a ruby filter.

hsiuming · November 23, 2017, 1:41am

Thanks for your reply.

Is there any method to solve question 2 ?

I try to write three out set in logstash conf file or split the logstash into three cong file .

But the output always contain three Binary in one data view on kibana.

I can not split it to three message on kibana.

magnusbaeck · November 23, 2017, 6:40am

Add the three values to an array, then use a split filter on that array.

hsiuming · November 24, 2017, 1:09am

Hello magnusbaeck

We now can split the event by binary rows.

The logstash config is :

if [message]=="XXXXX"
{
ruby {
code => "
event.set('[Binary1]', event.get('[event_data][Binary]')[0..7])
event.set('[Binary2]', event.get('[event_data][Binary]')[8..15])
event.set('[Binary3]', event.get('[event_data][Binary]')[16..23])

"
}

mutate {
    merge => ["Binary1","Binary2"]
}
mutate {
    merge => ["Binary1","Binary3"]
}

}

filter {

if [message]=="XXXXX"
{
split {
field => "Binary1"
remove_field => "Binary2"
remove_field => "Binary3"
}
}
}

Is there any method to add a id number or message to show the original binary location ?

I use the "add_field" function,but i don't know how to get the id number.

magnusbaeck · November 24, 2017, 6:34am

Is there any method to add a id number or message to show the original binary location ?

Id number of what? What do you mean by "binary location"?

If you show an example rather then describing what you want it'll probably be easier to understand what you mean.

hsiuming · November 24, 2017, 8:24am

Hello magnusbaeck

Following our logstash setting,we will get three event data in kibana.

The event data json is:

event1:

{
"_index": "wineventlog-2017-11-24",
"_type": "wineventlog",
"_id": "XXXXXXXXXXXXX",
"_version": 1,
"_score": null,
"_source": {
"computer_name": "XXXXXX",
"keywords": [
"XX"
],
"log_name": "Application",
"level": "info",
"record_number": "38868",
"event_data": {
"param1": "XXXXXXX",
"Binary": "000000000000000100000002"
},
"message": "XXXXXXX",
"type": "wineventlog",
"opcode": "info",
"tags": [
"beats_input_codec_plain_applied",
"_grokparsefailure"
],
"@timestamp": "2017-11-24T08:13:50.000Z",
"event_id": 1,
"Binary1": "00000000",
"@version": "1",
"beat": {
"name": "XXXXXX",
"hostname": "XXXXXX",
"version": "5.6.3"
},
"host": "XXXXXXX",
"source_name": "XXXXXXX"
},
"fields": {
"@timestamp": [
1511511230000
]
},
"highlight": {
"message": [
"(XXXXXXXXXXXXXXXXXXXXXXXX)"
]
},
"sort": [
1511511230000
]
}

event2:

{
"_index": "wineventlog-2017-11-24",
"_type": "wineventlog",
"_id": "XXXXXXXXXXXXX",
"_version": 1,
"_score": null,
"_source": {
"computer_name": "XXXXXX",
"keywords": [
"XX"
],
"log_name": "Application",
"level": "info",
"record_number": "38868",
"event_data": {
"param1": "XXXXXXX",
"Binary": "000000000000000100000002"
},
"message": "XXXXXXX",
"type": "wineventlog",
"opcode": "info",
"tags": [
"beats_input_codec_plain_applied",
"_grokparsefailure"
],
"@timestamp": "2017-11-24T08:13:50.000Z",
"event_id": 1,
"Binary1": "00000001",
"@version": "1",
"beat": {
"name": "XXXXXX",
"hostname": "XXXXXX",
"version": "5.6.3"
},
"host": "XXXXXXX",
"source_name": "XXXXXXX"
},
"fields": {
"@timestamp": [
1511511230000
]
},
"highlight": {
"message": [
"(XXXXXXXXXXXXXXXXXXXXXXXX)"
]
},
"sort": [
1511511230000
]
}

event3:

{
"_index": "wineventlog-2017-11-24",
"_type": "wineventlog",
"_id": "XXXXXXXXXXXXX",
"_version": 1,
"_score": null,
"_source": {
"computer_name": "XXXXXX",
"keywords": [
"XX"
],
"log_name": "Application",
"level": "info",
"record_number": "38868",
"event_data": {
"param1": "XXXXXXX",
"Binary": "000000000000000100000002"
},
"message": "XXXXXXX",
"type": "wineventlog",
"opcode": "info",
"tags": [
"beats_input_codec_plain_applied",
"_grokparsefailure"
],
"@timestamp": "2017-11-24T08:13:50.000Z",
"event_id": 1,
"Binary1": "00000002",
"@version": "1",
"beat": {
"name": "XXXXXX",
"hostname": "XXXXXX",
"version": "5.6.3"
},
"host": "XXXXXXX",
"source_name": "XXXXXXX"
},
"fields": {
"@timestamp": [
1511511230000
]
},
"highlight": {
"message": [
"(XXXXXXXXXXXXXXXXXXXXXXXX)"
]
},
"sort": [
1511511230000
]
}

Now we want to add a filed in each event:

event1:

{
"_index": "wineventlog-2017-11-24",
"_type": "wineventlog",
"_id": "XXXXXXXXXXXXX",
"_version": 1,
"_score": null,
"_source": {
"computer_name": "XXXXXX",
"keywords": [
"XX"
],
"log_name": "Application",
"level": "info",
"record_number": "38868",
"event_data": {
"param1": "XXXXXXX",
"Binary": "000000000000000100000002"
},
"message": "XXXXXXX",
"type": "wineventlog",
"opcode": "info",
"tags": [
"beats_input_codec_plain_applied",
"_grokparsefailure"
],
"@timestamp": "2017-11-24T08:13:50.000Z",
"event_id": 1,
"Binary1": "00000000",
"location": "1",
"@version": "1",
"beat": {
"name": "XXXXXX",
"hostname": "XXXXXX",
"version": "5.6.3"
},
"host": "XXXXXXX",
"source_name": "XXXXXXX"
},
"fields": {
"@timestamp": [
1511511230000
]
},
"highlight": {
"message": [
"(XXXXXXXXXXXXXXXXXXXXXXXX)"
]
},
"sort": [
1511511230000
]
}

event2:

{
"_index": "wineventlog-2017-11-24",
"_type": "wineventlog",
"_id": "XXXXXXXXXXXXX",
"_version": 1,
"_score": null,
"_source": {
"computer_name": "XXXXXX",
"keywords": [
"XX"
],
"log_name": "Application",
"level": "info",
"record_number": "38868",
"event_data": {
"param1": "XXXXXXX",
"Binary": "000000000000000100000002"
},
"message": "XXXXXXX",
"type": "wineventlog",
"opcode": "info",
"tags": [
"beats_input_codec_plain_applied",
"_grokparsefailure"
],
"@timestamp": "2017-11-24T08:13:50.000Z",
"event_id": 1,
"Binary1": "00000001",
"location": "2",
"@version": "1",
"beat": {
"name": "XXXXXX",
"hostname": "XXXXXX",
"version": "5.6.3"
},
"host": "XXXXXXX",
"source_name": "XXXXXXX"
},
"fields": {
"@timestamp": [
1511511230000
]
},
"highlight": {
"message": [
"(XXXXXXXXXXXXXXXXXXXXXXXX)"
]
},
"sort": [
1511511230000
]
}

event3:

{
"_index": "wineventlog-2017-11-24",
"_type": "wineventlog",
"_id": "XXXXXXXXXXXXX",
"_version": 1,
"_score": null,
"_source": {
"computer_name": "XXXXXX",
"keywords": [
"XX"
],
"log_name": "Application",
"level": "info",
"record_number": "38868",
"event_data": {
"param1": "XXXXXXX",
"Binary": "000000000000000100000002"
},
"message": "XXXXXXX",
"type": "wineventlog",
"opcode": "info",
"tags": [
"beats_input_codec_plain_applied",
"_grokparsefailure"
],
"@timestamp": "2017-11-24T08:13:50.000Z",
"event_id": 1,
"Binary1": "00000002",
"location": "3",
"@version": "1",
"beat": {
"name": "XXXXXX",
"hostname": "XXXXXX",
"version": "5.6.3"
},
"host": "XXXXXXX",
"source_name": "XXXXXXX"
},
"fields": {
"@timestamp": [
1511511230000
]
},
"highlight": {
"message": [
"(XXXXXXXXXXXXXXXXXXXXXXXX)"
]
},
"sort": [
1511511230000
]
}

Is it possible?

magnusbaeck · November 27, 2017, 6:39am

Okay. So right now you build up an array that look like this:

["000000000", "1111111111111", "2222222222222"]

Change your ruby filter to produce this instead:

[{"binary": "000000000", "location": "1", {"binary": "1111111111111", "location": "2"}, {"binary": "2222222222222", "location": "3"}]

That way, the field in question will after the split contain an object. The subfields binary and location can then be moved into the top level of the event. The ruby code could look like this:

event.set('binary', 
    [
        {"binary" => event.get('[event_data][Binary]')[0..7]), "location" => "1"},
        {"binary" => event.get('[event_data][Binary]')[8..15]), "location" => "2"},
        {"binary" => event.get('[event_data][Binary]')[16..23]), "location" => "3"}
    ])

hsiuming · November 27, 2017, 8:59am

Hello magnusbaeck

I change the filter to :

filter {

grok {
match => [ "message", "%{COMMONAPACHELOG}" ]
}
if [message]=="XXXXX"
{
ruby {
code => "
event.set('binary',
[
{"binary" => event.get('[event_data][Binary]')[0..7]), "location" => "1"},
{"binary" => event.get('[event_data][Binary]')[8..15]), "location" => "2"},
{"binary" => event.get('[event_data][Binary]')[16..23]), "location" => "3"}
])

"
}
}

}

filter {

if [message]=="XXXXX"
{
split {
field => "binary"
}
}

}

The logstash event log show:

[ERROR][logstash.agent ] Cannot create pipeline {:reason=>"Expected one of #, {, } at line 18, column 11 (byte 207) after filter {\n\n grok {\n match => [ "message", "%{COMMONAPACHELOG}" ]\n }\nif [message]=="(INTEGRATION)"\n{\nruby {\n code => "\nevent.set('binary', \n [\n {""}

Is there a issue about event.set form?

The version of logstash is 5.6.4 on ubuntu.

magnusbaeck · November 27, 2017, 9:07am

You can't use double quotes inside the ruby code block. Replace them with single quotes.

hsiuming · November 27, 2017, 9:24am

Hello magnusbaeck

The filter is :

filter {

grok {
match => [ "message", "%{COMMONAPACHELOG}" ]
}
if [message]=="XXXXX"
{
ruby {
code => "
event.set('binary',
[
{'binary' => event.get('[event_data][Binary]')[0..7]), 'location' => '1'},
{'binary' => event.get('[event_data][Binary]')[8..15]), 'location' => '2'},
{'binary' => event.get('[event_data][Binary]')[16..23]), 'location' => '3'}
])

"
}

split {
    field => "binary"
}

}

The event log not have error message.

But I can not see the data in kibana.

The logstash seems to not convert the data.

magnusbaeck · November 27, 2017, 12:39pm

Always use a stdout { codec => rubydebug } output while you're debugging. Once you're happy with what the events look like you can enable the elasticsearch output again.

hsiuming · November 28, 2017, 8:44am

Hello magnusbaeck

Finally I set the logstash config:

ruby {
code => "
event.set('[Binary1]', event.get('[event_data][Binary]')[0..7]+'_1')
event.set('[Binary2]', event.get('[event_data][Binary]')[8..15]+'_2')
event.set('[Binary3]', event.get('[event_data][Binary]')[16..23]+'_3')

"
}

And then use kibana Scripted fields to do it.

Very thanks for your help.

system · December 26, 2017, 8:44am

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

Topic		Replies	Views
Event.remove not working Logstash	5	3260	June 29, 2018
Need some help with Ruby event.set / get to use multiplication in config Logstash	3	5962	February 18, 2019
Help with Ruby code in Logstash Logstash	6	1393	September 4, 2018
Best practise method to split events and perform filtering to those events Logstash	2	487	December 27, 2016
Logstash creates switch case using ruby code Logstash	7	5313	September 2, 2019

Logstash event ruby code

Related topics