Forum Discussion

gof5go_289559's avatar
gof5go_289559
Icon for Nimbostratus rankNimbostratus
Apr 27, 2017

Health monitor to parse Nagios XI REST API call

I am needing to check Linux service status (crond) to determine if a pool member is healthy.

Initially, I started down the path to use an external monitor to run a perl script to do this, but have a bit to set up to allow the f5 to SSH with keys to all of our hosts.

So then I found that Nagios XI (for which there is already and agent set up on all of our hosts) has a REST API. The results returned with a curl request contain many attributes all in one line. For example:

> curl -k --silent "https://nagiosserver.local/nagiosxi/api/v1/objects/servicestatus?host_name=host5&service_description=crond&apikey=[api-key-here]

{"servicestatuslist":{"recordcount":"1","servicestatus":{"@attributes":{"id":"390234"},"instance_id":"1","service_id":"2214","host_id":"2208","host_name":"host5","host_alias":"host5","name":"crond","host_display_name":"","host_address":"10.10.10.5","display_name":"crond","status_update_time":"2017-04-27 12:27:18","status_text":"crond (pid  3398) is running...","status_text_long":"","current_state":"0","performance_data":"","should_be_scheduled":"1","check_type":"0","last_state_change":"2017-04-27 09:04:08","last_hard_state_change":"2017-04-27 09:04:08","last_hard_state":"0","last_time_ok":"2017-04-27 12:22:48","last_time_warning":"1969-12-31 19:00:00","last_time_critical":"1969-12-31 19:00:00","last_time_unknown":"1969-12-31 19:00:00","last_notification":"1969-12-31 19:00:00","next_notification":"1969-12-31 19:00:00","no_more_notifications":"0","acknowledgement_type":"0","current_notification_number":"0","process_performance_data":"1","obsess_over_service":"1","event_handler_enabled":"1","modified_service_attributes":"0","event_handler":"","check_command":"check_nrpe!check_init_service!-a 'crond'!!!!!!","normal_check_interval":"5","retry_check_interval":"1","check_timeperiod_id":"115","icon_image":"","icon_image_alt":"","has_been_checked":"1","current_check_attempt":"1","max_check_attempts":"5","last_check":"2017-04-27 12:22:48","next_check":"2017-04-27 12:27:46","state_type":"1","notifications_enabled":"0","problem_acknowledged":"0","flap_detection_enabled":"1","is_flapping":"0","percent_state_change":"0","latency":"0","execution_time":"0.08153","scheduled_downtime_depth":"0","passive_checks_enabled":"1","active_checks_enabled":"1"}}}

I really only need to check that "current_state" = 0 or that "status_text" contains "is running". The following greps out the current_state of "0" which would mean my pool member should be marked UP.

curl -k --silent "https://nagiosserver.local/nagiosxi/api/v1/objects/servicestatus?host_name=host5&service_description=crond&apikey=[api-key-here]" | grep -Po 'current_state":"\K.*?(?=",)'
O

So, how to make this a health monitor? Is it possible to do this using the built-in http/https monitor types or is using an external monitor the only way since the URL used is always my nagios server's and not the pool member's?

Many thanks!

application delivery
LTM

1 Reply

Sort By
  • gof5go_289559's avatar
    gof5go_289559
    Icon for Nimbostratus rankNimbostratus
    Apr 28, 2017

    I ended up creating an external file monitor with the following. Still need to fix the "-k" requirement on the curl command...

    In the LTM health monitor, add a variable with the Name of HOST_NAME and the Value of the host's name as seen by Nagios.

    This could easily be re-purposed to check any service's status in Nagios...

    !/bin/sh

 This script expects the following Name/Value in the LTM monitor and no Arguments are needed
  HOST_NAME  = the monitored host name in Nagios

 remove IPv6/IPv4 compatibility prefix (LTM passes addresses in IPv6 format)
IP=`echo ${1} | sed 's/::ffff://'`
PORT=${2}

PIDFILE="/var/run/`basename ${0}`.${IP}_${PORT}.pid"
 kill of the last instance of this monitor if hung and log current pid
if [ -f $PIDFILE ]
then
   echo "EAV exceeded runtime needed to kill ${IP}:${PORT}" | logger -p local0.error
   kill -9 `cat $PIDFILE` > /dev/null 2>&1
fi
echo "$$" > $PIDFILE

  send request & check for "current_state:0"
    curl -kfNs "https://nagiosserver.local/nagiosxi/api/v1/objects/servicestatus?host_name=${HOST_NAME}&service_description=crond&apikey=[api-key-here]" | grep -i "\"current_state\":\"0\"" 2>&1 > /dev/null

     mark node UP if true
    if [ $? -eq 0 ]
    then
        rm -f $PIDFILE
        echo "UP"
    else
        rm -f $PIDFILE
    fi