NEET (N agios E xternal E xperiment T ests), is a tool to send experiment tests via "MSG" (Message BUS) to a specific QUEUE. Nagios reads from the queue and publishes the information.

NEET (Nagios External Experiment Tests)

Neet is written in python. It contains 3 files:

  • neet.cfg

Usage ""

In order to use NEET, you have to follow these steps:

1. Run . The script reads the information regarding the experiment tests from file. The path to the input file must be passed as first argument to the script. The input file must contain a section with the metrics used in the input file itself. The metrics must be described in a dictionary structure and for each field the user has to specify the position of the field in the input file, i.e.::

show hide

_metrics  = {
            'hostName' : {'metric_description': " Host Name",
                                'field_position'   : 0,
            'QueueName' : {'metric_description': "Queue Name",
                                'field_position'   : 1,
            'spaceTokenName' : {'metric_description': "Space Token Name",
                                'field_position'   : 2,
            'status' : {'status': "Status in string format",
                                'field_position'   : 3,
            'metricStatus' : {'metric_description': "Status in number format",
                                'field_position'   : 4,
            'jobId' :  {'metric_description': "Job ID",
                                'field_position'   : 5,
            'AthenaExitCode' : {'metric_description': "Athena Exit Code",
                                'field_position'   : 6,
            'metricName' : {'metric_description': "......",
                                'field_position'   : 7,
            'dataSetName' : {'metric_description': "......",
                                'field_position'   : 8,
            'startTime' : {'metric_description': "start time of the Job",
                                'field_position'   : 9,
            'endTime' : {'metric_description': "end time of the Job",
                                'field_position'   : 10,

2. Create a new dictionary with the specific metric for Nagios, for example:

metricName: org.atlas.WN-gangarobot_wms-atlas
metricStatus: OK
summaryData: mc09_7TeV.105193  
detailsData: mc09_7TeV.105193.CompHepPythia....

3. call the method send of exp_test_result_publisher to dispatch the information into MSG

Usage ""

This script, is only used to publish information to Nagios,

The method format_Message takes the metrics dictionary as first parameter and uses it to create a message in "WLCG format" which can be read by Nagios. The message is sent via MSG to a dedicated QUEUE where Nagios is listening.

  send ('hostName: '+_metrics ['hostName'] +'\n'+
        'metricName: '+_metrics ['metricName']+'\n'+
        'metricStatus: '+_metrics ['metricStatus']+'\n'+
        'timestamp: '+_metrics ['timeStamp']+'\n'+
        'summaryData: '+_metrics ['summaryData']+'\n'+
        'detailsData: '+_metrics ['detailsData']+'EOT\n') 

  conn.send(message,destination=QUEUE, ack='auto')
Usage "neet.cfg"

The user must use this file to define the parameters for the connection to MSG using the stomp protocol.

PORT: 6163
QUEUE: /queue/grid.probe.metricOutput.EGEE.bece5c857f68263c125890256fa8833c

Example of usage

Now, Neet works to publish GangaRobot Experiment Test result , it's running with the user ddmusr03 in:


There is a script file,, running hourly, that downloads a GangaRobot Experiment files (type LatestResult*) and gives one file at a time to

$HOME/public/nagios_publisher/ -name 'LatestResults*'  -exec  $HOME/public/nagios_publisher/ {} \; 

How to configure NAGIOS box to accept external tests

Client side

The lightweight library used to send messages via stomp protocol to a Message Bus requires a small configuration client side (referring to as neet.conf). Once the formatted input file for the parser, one has to insure that the publisher script points to the right configuration file:

this is the relevant part of the file to be modified"/afs/")
and then editing this file cfg file one sets the destination message bus endpoints.

[lxplus406] ~/scratch0/grid/conditionDB/publisher $ cat /afs/ 
# Neet config file

PORT: 6163
#QUEUE: /queue/grid.probe.LHCb-externals
QUEUE: /queue/grid.probe.metricOutput.EGEE.sam-lhcb_cern_ch

Server side

The publication of well formatted test results in the Message Bus via NEET is a trivial operations; the configuration of the Nagios box to consume these results is also very simple but a bit of care must be payed. These tests are passive checks but they do not have parents active checks in Nagios. A clean solution proposed in this document is to define in the ncg.conf file a profile ad hoc where generic VO test can be later defined in the We show in this example the ncg.conf file used by LHCb:
$ cat /etc/ncg/ncg.con
      VO_FQAN = /lhcb/Role=production
      VO_FQAN = lhcb
      VO_FQAN = /lhcb/Role=pilot

We have defined two profiles:

  • VO_PROD: used to define all metrics running within Nagios with Role=production
  • VO_PILOT: used to define all metrics running within Nagios with Role=pilot
  • VO_EXTERNALS: used to define all metrics running outside Nagios with undefined FQAN used to be submitted.
Please note that the metricName of your external tests (as will be known to Nagios) will be the metric name + "-lhcb" appended at the end. In your clients please insure that metricName field reflects exactly this convention otherwise Nagios will not publish results. Now we have to inform Nagios about these new external metrics and this is almost identical of the definition of any other test. Add the following lines in your for each of the metrics that you want to publish. Please note here the parent is org.sam.CE-JobState just because we wanted to publish this WN specific tests with the rest of the WN tests submitted via Nagios. But this is not compulsory. Took exactly this example and change the name of your test from org.lhcb.WN-CondDB to whatever you want.
#### external submission via NEET
$WLCG_SERVICE->{'org.lhcb.WN-CondDB'}->{flags}->{PASSIVE} = 1;
$WLCG_SERVICE->{'org.lhcb.WN-CondDB'}->{parent} = "org.sam.CE-JobState";
$WLCG_SERVICE->{'org.lhcb.WN-CondDB'}->{flags}->{VO} = 1;
$WLCG_SERVICE->{'org.lhcb.WN-CondDB'}->{flags}->{OBSESS} = 1;
$WLCG_SERVICE->{'org.lhcb.WN-CondDB'}->{metricset} = 'org.lhcb.WN';

At this point you have to inform Nagios that this metrics belongs to the VO_EXTERNAL profile for the service CE. This is trivially done as described in the relevant snipped of the

### external publisher for VO_external


At this point just re-configure your nagios box and restart all services as reported in the following procedure.

-- AleDiGGi - 16-Aug-2010

Topic attachments
I Attachment History Action Size Date Who Comment
PDFpdf DavideM-NagiosPub_17082010.pdf r1 manage 3608.4 K 2010-08-19 - 10:48 UnknownUser  
PowerPointppt DavideM-NagiosPub_17082010.ppt r1 manage 428.5 K 2010-08-19 - 10:46 UnknownUser  
Edit | Attach | Watch | Print version | History: r11 < r10 < r9 < r8 < r7 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r11 - 2010-11-29 - unknown
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2023 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback