Nagios Support Forum

Posted: **Tue Feb 16, 2016 6:49 pm**

Hi I'm currently trying to import from a file, I've attempted to follow the steps from the 'source setup' within Nagios but have run into a problem.

Firstly my experience using Linux is pretty much non existent, I've downloaded the shipper.py script and installed netcat.

When attempting to ship a log file as a test I get these errors.

Code: Select all

[root@xxxxxxxx ~]# cat /var/log/httpd/access_log-20160131 | python shipper.py program:apache_access | nc xxx.xxx.xx.xxx 2056
[Errno 32] Broken pipe
Traceback (most recent call last):
  File "shipper.py", line 242, in <module>
    main()
  File "shipper.py", line 237, in main
    process_stream(sys.stdin, message)
  File "shipper.py", line 217, in process_stream
    print json.dumps(message)
IOError: [Errno 32] Broken pipe

Where do I go from here, is this an issue with the code in shipper.py or have I messed up somewhere?

Any help would be great!, thanks.

Posted: **Wed Feb 17, 2016 10:32 am**

Where do I go from here, is this an issue with the code in shipper.py or have I messed up somewhere?

Let's take this troubleshooting process in digestible chunks - first, we'll see if you can ship plaintext to Nagios Log Server.

Code: Select all

echo 'test words' | nc xxx.xxx.xx.xxx 2056

If the above arrives properly, you know that you're fine to move forward and attempt to send a file - let me know how the above works out.

Once the log arrives in Nagios Log Server, you will be able to see it in the 'Dashboards' section, you can query for "test words."

Posted: **Thu Feb 18, 2016 6:45 pm**

Thanks Jolson,

Basically troubleshooting with the test words code you gave me, I figured I made a rookie mistake in using the wrong IP (external instead of internal address).

My next question is, as I am importing logs the @timestamp field won't accurately show for the imported logs. @timestamp field will be of when I imported the logs.

What have i tried doing is I've created an addition field called I_timestamp with a grok filter which grabs the timestamp from the log, how do I parse this as a timestamp? currently it is a string and I'd like to be able to graph it.

Grok filter below:

Code: Select all

if [host] =='xxx.xxx.xx.xxx' {
    grok {
        match => ['message', '%{TIMESTAMP_ISO8601:I_timestamp}']
    }
}

timestamp_import.PNG

An example of the beginning of my log lines.

Code: Select all

2016-02-18 11:11:00,150 INFO  [RequestDetailValve.logMessages:100 852

Error message on the histogram.

Code: Select all

 FacetPhaseExecutionException[Facet [0]: (key) field [I_timestamp] not found]

Could you please guide me on the best way of implementing this.

Thanks!

Posted: **Fri Feb 19, 2016 11:09 am**

Could you break the timestamp down further using additional grok filters to give you the data that you want?

For instance:

Code: Select all

2016-02-18 11:11:00

Code: Select all

%{DATE:date} %{TIME:time}

Code: Select all

{
  "date": [
    [
      "16-02-18"
    ]
  ],
  "DATE_US": [
    [
      null
    ]
  ],
  "MONTHNUM": [
    [
      null,
      "02"
    ]
  ],
  "MONTHDAY": [
    [
      null,
      "16"
    ]
  ],
  "YEAR": [
    [
      null,
      "18"
    ]
  ],
  "DATE_EU": [
    [
      "16-02-18"
    ]
  ],
  "time": [
    [
      "11:11:00"
    ]
  ],
  "HOUR": [
    [
      "11"
    ]
  ],
  "MINUTE": [
    [
      "11"
    ]
  ],
  "SECOND": [
    [
      "00"
    ]
  ]
}

I'm not sure if the ,150 was relevant, but this could be a way to do it.

Posted: **Sun Feb 21, 2016 6:48 pm**

Thanks for the response, I may have been unclear with what I was trying to accomplish.

Will breaking the data down further allow me to graph the data in a histogram?

As it is an import the @timestamp field will all mostly be identical and when graphing in an histogram will not give an accurate representation of when my log messages appear.

The @timestamp field format is : YYYY-MM-DDTHH-mm:ss.SSSZ. So I figured something similar (my I_timestamp) would work but it is parsing as a string instead of what @timestamp is and the histogram is not working

Code: Select all

@timestamp: 2016-02-18T23:31:55.985Z

Code: Select all

I_timestamp: 2016-02-18 11:11:00,150

Posted: **Mon Feb 22, 2016 1:56 pm**

Is this histogram something you're using outside of NLS?

Posted: **Mon Feb 22, 2016 10:31 pm**

No, this is the histogram within Nagios Log Server.

histogram.PNG

The error on the histogram.

histogram_error.PNG

Posted: **Tue Feb 23, 2016 2:25 pm**

Is this a custom dashboard, or one of the default dashboards? I'm trying to do some research on this, and it sounds like there may be some issues with certain types of dashboards. If you're interested, here is the thread I'm looking at.

Posted: **Tue Feb 23, 2016 8:10 pm**

Thanks for the response.

This is a default dashboard, after reading around some threads I think the issue is that my new field 'im_timestamp' is being parsed as a string and the histogram doesn't like it.

I tried parsing it as a date in the grok filter with the code below:

Code: Select all

    grok {
        match => ['message', '%{TIMESTAMP_ISO8601:im_timestamp:date}']
    }

It is still a string though. Do you any suggestions on this?

im_timestamp(string).PNG

Posted: **Wed Feb 24, 2016 2:55 pm**

Yes - you'll have to make use of the 'date' filter, like so:

Code: Select all

date {
match => [ "im_timestamp", "MMM dd YYY HH:mm:ss",
          "MMM  d YYY HH:mm:ss", "ISO8601" ]
}

Of course you can read more about the date filter here: https://www.elastic.co/guide/en/logstas ... -date.html

Nagios Support Forum

Importing from file - assistance please

Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please

Re: Importing from file - assistance please