EventHandler fires too often on SNMP-TrapDirector

Icinga 2
1

I am running SNMP-TrapDirector to handle Interface Up/Down traps.
The handler sets the state of a switch’s service definition to OK (Up) or WARNING (Down).

Whenever a port comes up, we run an automation comparing the MAC on the switch port with the one in our documentation (Netbox). We created a simple EventCommand for this which calls a Webhook in an Ansible automation.

The Request should only fire if:

  • service_state = OK
  • service_state_type = HARD
  • service_last_state != OK
  • service_long_output ends in " Up"

Unfortunatly, our Backend always receives two identical automation requests :frowning:

grafik

I guess what I am looking for is a state-machine for the service state - ideally I want to filter in the first “Up” event.

Thanks already!
Alex

2

It’s a while since I did anything with the event handler.

If I remember correctly, I also passed the previous state and it worked.
Can you post the config of your EventCommand and the filter code at the top in the script you call?

3

Thank you for you quick reply.

This is what I use:

object EventCommand "network_port_event" {
  import "plugin-event-command"
  command = [ PluginDir + "/network-port-up-down-handler.sh" ]
  arguments = {
        "--state" = "$service.state$"
        "--last-state" = "$service.last_state$"
        "--state-type" = "$service.state_type$"
        "--last-state-type" = "$service.last_state_type$"
        "--output" = "$service.output$"
        "--host" = "$host.name$"
        "--ip" = "$host.address$"
        "--service" = "$service.name$"
  }
}

network-port-up-down-handler.sh

#!/bin/bash

STATE=“”
STATE_TYPE=“”
LAST_STATE=“”
LAST_STATE_TYPE=“”
OUT=“”
IP=“”
SERVICE=“”

while [[ “$#” -gt 0 ]] ; do
case $1 in
–state)
STATE=$2 ; shift ;;
–state-type)
STATE_TYPE=$2 ; shift ;;
–last-state)
LAST_STATE=$2 ; shift ;;
–last-state-type)
LAST_STATE_TYPE=$2 ; shift ;;
–output)
OUT=$2 ; shift ;;
–host)
HOST=$2 ; shift ;;
–ip)
IP=$2 ; shift ;;
–service)
SERVICE=$2 ; shift ;;
*)
echo “Unknown parameter: $1” ; exit 1 ;;
esac
shift
done

if [[ “$STATE” = “OK” && “$STATE_TYPE” = “HARD” && “$LAST_STATE” != “OK” ]]; then
T=${OUT## Interface }
INTERFACE=${T%% Up}
if [ “${OUT##* }” = “Up” ] ; then
/usr/bin/curl -s https://ansible/api/integrations/test123 -X POST --data-raw “{“device_name”:”${HOST}“,“interface_name”:”${INTERFACE}“}” -H “Content-Type: application/javascript”
fi
fi

exit 0

4

LGTM… next log into a file something like datetime var1 var2 … to track down, what happens.

I tired to find my old config and code but it was in the last instance I can’t find right now.

5

Great idea about writing a debug log :slight_smile:

It’s indeed a problem with the switch sending unsolicited traps. I should have checked before…

Sorry for the noise!