prev. article next article
shinken

shinken 1.2.4 on ubuntu desktop 12.10 ii

8 Mar 2013

copied from shinken-monitoring.org

motivation

in this posting i want to enhance the shinken usage with a few fixes

writing custom service checks for shinken

there exsists a nice documentation for nagios [1] already. i didn’t find much in the shinken wiki, so here is what i did.

the test script in /usr/local/shinken/libexec/check_test

#! /bin/bash
#
# js@lastlog.de
# 05/03/2013
#
#  This Nagios test plugin was created to demonstrate how it is integrated into shinken
#

PROGNAME=`basename $0`
PROGPATH=`echo $0 | sed -e 's,[\\/][^\\/][^\\/]*$,,'`
REVISION="0.0.1"

. $PROGPATH/utils.sh

print_usage() {
  echo "Usage:"
  echo "  $PROGNAME --help"
  echo "  $PROGNAME --test"
  echo "  $PROGNAME --version"
}

print_help() {
  print_revision $PROGNAME $REVISION
  echo ""
  print_usage
  echo ""
  echo "Nagios test plugin"
  echo ""
  echo "--test"
  echo "   Perform a test; in this implementation it either returns: STATE_OK, STATE_CRITICAL, STATE_WARNING or STATE_UNKNOWN "
  echo "--help"
  echo "   Print this help screen"
  echo "--version"
  echo "   Print version and license information"
  echo ""
  support
}

# Information options
case "$1" in
--help)
                print_help
    exit $STATE_OK
    ;;
-h)
                print_help
    exit $STATE_OK
    ;;
--version)
                print_revision $PROGNAME $REVISION
    exit $STATE_OK
    ;;
-V)
                print_revision $PROGNAME $REVISION
    exit $STATE_OK
    ;;
--test)
    # STATE_OK
    # STATE_WARNING
    # STATE_CRITICAL
    # STATE_UNKNOWN
    # STATE_DEPENDENT
    # $(($RANDOM%3))
    case $(($RANDOM%5)) in
      0)
          echo "STATE_OK - $(date)"
          exit $STATE_OK
      ;;
      1)
          echo "STATE_WARNING - $(date)"
          exit $STATE_WARNING
      ;;
      2)
          echo "STATE_CRITICAL - $(date)"
          exit $STATE_CRITICAL
      ;;
      3)
          echo "STATE_DEPENDENT - $(date)"
          exit $STATE_DEPENDENT
      ;;
      *)
          echo "STATE_UNKNOWN - $(date)"
          exit $STATE_UNKNOWN
      ;;
    esac
    ;;
*)
    print_usage
    exit $STATE_UNKNOWN
esac

also check:

chown shinken:shinken check_test 
chmod 0755 check_test 

execute it on a shell, it should return:

root@monitoring-vm:/usr/local/shinken/libexec# ./check_test --test
STATE_CRITICAL - Thu Mar  7 12:34:19 CET 2013
root@monitoring-vm:/usr/local/shinken/libexec# ./check_test --test
STATE_WARNING - Thu Mar  7 12:44:23 CET 2013
root@monitoring-vm:/usr/local/shinken/libexec# ./check_test --test
STATE_UNKNOWN - Thu Mar  7 12:44:23 CET 2013
root@monitoring-vm:/usr/local/shinken/libexec# ./check_test --test
STATE_DEPENDENT - Thu Mar  7 12:44:24 CET 2013
root@monitoring-vm:/usr/local/shinken/libexec# ./check_test --test
STATE_OK - Thu Mar  7 12:44:25 CET 2013
root@monitoring-vm:/usr/local/shinken/libexec# ./check_test --test
STATE_WARNING - Thu Mar  7 12:44:26 CET 2013

/usr/local/shinken/etc/commands.cfg

append this to commands.cfg

define command {
    command_name    check_test
    command_line    $PLUGINSDIR$/check_test $ARG1$
}

/usr/local/shinken/etc/services/services.cfg

define service{
       use generic-service
       service_description CustomTest
       host_name lastlog.de
       check_command check_test!"--test"
}

note: use generic-service means that our custom service inherits all attributes from the template generic-service defined in /usr/local/shinken/etc/templates.cfg

testing the new rules

  1. restart shinken to test if the new rules work in case of an configuration error, shinken will bail out with:

    /etc/init.d/shinken restart
    Doing config check
     * full result is in /tmp/shinken_checkconfig_result
     * ConfigCheck failed: Configuration is incorrect, sorry, I bail out
       ...fail!
  2. have a look at the webinterface, for every test invocation the status is very likely to change between: STATE_OK, STATE_CRITICAL, STATE_WARNING or STATE_UNKNOWN

shinken webinterface checks

if everything is working you should see something like this:

note: [4] might be a good introduction into advanced dependencies of service checks!

thanks for the help to Issif#shinken@irc.freenode.net and [1]

using thruk with shinken

deploy thruk [3] and disable shinken webinterface

installation

wget http://www.thruk.org/files/pkg/v1.64-2/ubuntu12.10/amd64/thruk_1.64-2_ubuntu12.10_amd64.deb
dpkg -i thruk_1.64-2_ubuntu12.10_amd64.deb
apt-get -f install    # as listed in the manual at [5]

enable livestatus for shinken

shinken has livestatus (127.0.0.1:50000, a python process) already enabled by default. so we need to edit thruk_local.cfg

netstat -tulpen
...
tcp        0      0 0.0.0.0:50000           0.0.0.0:*               LISTEN      999        35960       6918/python

/etc/thruk/thruk_local.cfg

  1. before making this change i had this problem:

    less /var/log/apache2/error.log
    ...
    [Thu Mar 07 15:03:35 2013] [notice] child pid 17670 exit signal Segmentation fault (11)
    [Thu Mar 07 15:03:35 2013] [notice] child pid 17671 exit signal Segmentation fault (11)
    [Thu Mar 07 15:03:35 2013] [notice] child pid 17672 exit signal Segmentation fault (11)
    [Thu Mar 07 15:03:35 2013] [notice] child pid 17673 exit signal Segmentation fault (11)
    ...
  2. this few lines need to be uncommented (so remove all the # symbols)

    ############################################
    # put your own settings into this file
    # settings from this file will override
    # those from the thruk.conf
    ############################################
    
    ######################################
    # Backend Configuration, enter your backends here
    <Component Thruk::Backend>
        <peer>
            name   = External Shinken
            type   = livestatus
            <options>
                peer    = 127.0.0.1:50000
           </options>
        </peer>
    </Component>
  3. restart thruk and apache2

    /etc/init.d/thruk restart
    /etc/init.d/apache2 restart
  4. finally visit: http://localhost/thruk: user thrukadmin with password thrukadmin

make thruk work with RRD pnp4nagios graphs

to make this work i used some parts of the documentation [7] but it it required some serious debugging and my current result looks different to what can be found on http://demo.thruk.org.

  1. basically append the action_url to the generic-service record in the templates.cfg file as shown below:

    # Generic service definition template - This is NOT a real service, just a template!
    define service{
         name                            generic-service         ; The 'name' of this service template
         active_checks_enabled           1                       ; Active service checks are enabled
         passive_checks_enabled          1                       ; Passive service checks are enabled/accepted
    
         # Check part
         # By default, there is no check_command here
         check_interval           5                      ; Check the service every 5 minutes in normal state
         retry_interval           1                      ; Re-check the service every one minutes until a hard state can be determined
         max_check_attempts       2                      ; Re-check the service up to 3 times in order to determine its final (hard) state
         check_period             24x7                   ; The service can be checked at any time of the day
    
         # Notification part
         notifications_enabled           1                       ; Service notifications are enabled
         notification_options            w,u,c,r                 ; Send notifications about warning, unknown, critical, and recovery events
         notification_interval           60                      ; Re-notify about service problems every hour
         notification_period             24x7
         # If the contacts and contact_groups options are not set, it will notify host contacts instead
         # contact_groups                  admins
    
         # Advanced options. Change with care
         #event_handler_enabled           1
         # event_handler                 super_event_kill_everyone!DIE
         flap_detection_enabled          1                       ; Flap detection is enabled
         check_freshness                 0
         freshness_threshold             3600
         #stalking_options                w,c
         obsess_over_service             0
         #escalations                    ToLevel2
         process_perf_data               1                       ; Process perf data, like for PNP
         is_volatile                     0                       ; for log monitoring. See doc for more info about it
    
         # For the WebUI
         icon_set                         server ; can be database, disk, network_service, server
    
         register                        0
         action_url /pnp4nagios/graph?host=$HOSTNAME$&srv=$SERVICEDESC$' class='tips' rel='/pnp4nagios/index.php/popup?host=$HOSTNAME$&srv=$SERVICEDESC$
         }
  2. i also made both NPCDMOD changes both in shinken-specific.cfg.

results

security

  1. change the password for user thrukadmin

  2. pnp4nagios uses a password, make thruk aware of that, see pnp_export in [6]

  3. disable the shinken webinterface in /usr/local/shinken/etc/shinken-specific.cfg

    1. just remove the WebUI from the modules list below

      define broker {
        broker_name broker-1
        data_timeout 120
        check_interval 60
        modules Livestatus, WebUI, Simple-log, NPCDMOD
        port 7772
        manage_sub_realms 1
        spare 0
        timeout 3
        address localhost
        realm All
        max_check_attempts 3
        manage_arbiters 1
      }
    2. restart shinken

      /etc/init.d/shinken restart
    3. verify that the webinterface is not there anymore: http://localhost:7767

resources needed

it seems that thruk is quite memory hungry. in my current setup (13 hosts / 25 services) about 1gb ram minimum is required.

proxy shinken webinterface via vhost through apache2

put shinken webservice (port 7767) behind apache vhosts, thus setup a proxy to make the webservice available via port 80 orr 443, see [2] and even better [8] and [9].

modify default vhost in /etc/apache2/sites-available/default

i’ve added these lines to the configuration file:

    ...
    # prevent a forward proxy! 
    ProxyRequests off

    # User-Agent / browser identification is used from the original client
    # shinken will then return either the mobile or desktop version of the webpage!
    ProxyVia Off
    ProxyPreserveHost On 

    # since on ubuntu it is disabled by default, we have to reenable it here
    # i don't want to touch /etc/apache2/mods-enabled/proxy.conf
    <Proxy *>
    Order deny,allow
    Allow from all
    </Proxy>

    # prevent pnp4nagios from being reverse-proxied
    ProxyPass /pnp4nagios !
    ProxyPass / http://localhost:7767/
    ProxyPassReverse / http://localhost:7767/

</VirtualHost>

security

as outlined in [10] i think this configuration should be safe as it does not create an open forward proxy since there is ProxyRequests off configured. also the reverse proxy is configured to access only http://localhost:7767/.

this can be verified by this commands:

tcpdump src host 1.2.3.4 and dst port 80

first start telnet

telnet localhost 80

then issue this get (hit return 2 times): GET http://www.google.com HTTP/1.0

this is the trace:

# telnet localhost 80
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
GET http://www.google.com HTTP/1.0
 
HTTP/1.1 303 See Other
Date: Sat, 09 Mar 2013 23:35:11 GMT
Server: PasteWSGIServer/0.5 Python/2.6.5
Content-Length: 0
Content-Type: text/html; charset=UTF-8
Location: http://www.google.com/user/login
Vary: Accept-Encoding
Connection: close
 
Connection closed by foreign host.

note: there MUST BE no output in the tcpdump shell! i did enable ProxyRequests on for a short test, and then after restarting apache you actually have an open forward proxy which will cause much trouble! for the same request to google one will get a lot of traffic in the tcpdump command then.

thanks VERY MUCH to konkat, Unbelieve and Humbedooh from httpd#irc.freenode.org!

note: one thing that i still don’t understand is that konkat got HTTP/1.1 403 Forbidden where i get a HTTP/1.1 303 See Other.

#links * [1] http://community.spiceworks.com/how_to/show/3773-creating-custom-nagios-plugins-scripts-in-bash * [2] http://www.shinken-monitoring.org/forum/index.php?topic=192.0 * [3] http://www.thruk.org/documentation.html * [4] http://www.shinken-monitoring.org/wiki/setup_advanced_dependencies_in_shinken * [6] http://www.thruk.org/documentation.html#_installation * [7] http://www.shinken-monitoring.org/wiki/use_with_pnp * [8] http://adolfomaltez.wordpress.com/2011/05/27/apache-as-a-reverse-proxy/ * [9] http://andriigrytsenko.net/2011/02/apache-as-reverse-proxy-for-https-server/ * [10] http://httpd.apache.org/docs/2.2/mod/mod_proxy.html#access