Ad Widget

Collapse

Monitoring Host by simple ping check - number of pings before notify

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • AdiM
    Junior Member
    • Mar 2018
    • 1

    #1

    Monitoring Host by simple ping check - number of pings before notify

    Hi Guys,

    I was implementing Zabbix recently and started with simple host checks -
    I'm monitoring external server (just for this example, i added server address 8.8.4.4) with the following configurations:
    -Added host to the hosts list
    -Added application named "ICMP Test"
    -Added item with the following configurations:
    Name: ICMP Test
    Type: Simple Check
    Key: icmpping[8.8.4.4,1,2000,32,2000]
    Host Interface: 8.8.4.4:10050
    Type of information: Numeric (unsigned)
    Update interval: 30s
    Custom Intervals: Flexible
    Show value: As is
    Applications: ICMP Test
    Populates host inventory field: None
    -Added trigger with the following configurations:
    Name: Ping dropping on Server1 for more then 20 seconds
    Severity: choose whatever you want
    Expression:
    {8.8.4.4:icmpping[8.8.4.4,1,2000,32,2000].last(20)}<1
    OK event generation: Expression
    PROBLEM event generation mode: Single
    OK event closes: all problems

    Now, my problem is - according to the configuration, Zabbix checking the external host every 30 seconds, in case the ping is not response, Zabbix system sending notification to me, but since i'm working in a place with unstable internet, it can happens a lot.
    my goal is that Zabbix will check the ping and only in case 4-5 pings are not responding in range of 10-15 second, the trigger will send the notification.

    can you help me change the trigger configurations?

    Thank you
    Adi Mahluf
  • tcilmo
    Senior Member
    • Nov 2016
    • 122

    #2
    Originally posted by AdiM
    Hi Guys,

    I was implementing Zabbix recently and started with simple host checks -
    I'm monitoring external server (just for this example, i added server address 8.8.4.4) with the following configurations:
    -Added host to the hosts list
    -Added application named "ICMP Test"
    -Added item with the following configurations:
    Name: ICMP Test
    Type: Simple Check
    Key: icmpping[8.8.4.4,1,2000,32,2000]
    Host Interface: 8.8.4.4:10050
    Type of information: Numeric (unsigned)
    Update interval: 30s
    Custom Intervals: Flexible
    Show value: As is
    Applications: ICMP Test
    Populates host inventory field: None
    -Added trigger with the following configurations:
    Name: Ping dropping on Server1 for more then 20 seconds
    Severity: choose whatever you want
    Expression:
    {8.8.4.4:icmpping[8.8.4.4,1,2000,32,2000].last(20)}<1
    OK event generation: Expression
    PROBLEM event generation mode: Single
    OK event closes: all problems

    Now, my problem is - according to the configuration, Zabbix checking the external host every 30 seconds, in case the ping is not response, Zabbix system sending notification to me, but since i'm working in a place with unstable internet, it can happens a lot.
    my goal is that Zabbix will check the ping and only in case 4-5 pings are not responding in range of 10-15 second, the trigger will send the notification.

    can you help me change the trigger configurations?

    Thank you
    Adi Mahluf
    I am not sure if that exact calculation is possible. However, defining triggers like this might help. Set the key to run 3 pings.

    Missed 1/3 Pings --> ({Tolerant Ping:icmpping.avg(#3)}<1) and ({Tolerant Ping:icmpping.avg(#3)}>0.5)
    Missed 2/3 Pings --> ({Tolerant Ping:icmpping.avg(#3)}<0.5) and ({Tolerant Ping:icmpping.avg(#3)}>0)
    Missed 3/3 Pings --> {Tolerant Ping:icmpping.avg(#3)}=0

    Comment

    Working...