Re: bug or user error using ping

Paul Theodoropoulos via This is the general mailing list for monit <[email protected]> Mon, 6 Nov 2023 11:12:40 -0800
Newsgroups gmane.comp.monitoring.monit.general
Message-ID <[email protected]>
My apologies, yes, I'm running Monit v 5.33.0. I'll follow up with support.

I'm still curious about the questions I posed regarding response 
time/timeout and their interactions.

On 11/6/2023 1:19 AM, M/Monit Support wrote:
> Hello Paul,
>
> which Monit version it is?(the ping response time requires Monit 5.33.0)
>
> Is it possible to get a ping network trace between Monit and the 
> monitored machine and send it to [email protected] once the problem 
> occurs, along with the Monit log?
>
> It may also help to run Monit in debug mode (it'll log all ping tests, 
> including the count, and response time).
>
> Best regards,
> The M/Monit team
>
>
>> On 5. 11. 2023, at 19:23, Paul Theodoropoulos via This is the general 
>> mailing list for monit <[email protected]> wrote:
>>
>> Greetings. I'm using monit to check a device on my local lan to ensure 
>> it's responsive. This is what I was previously using:
>>
>>     check host resolver2 address 192.168.1.13
>>     every 5 cycles
>>     if failed ping4
>>     then exec /usr/local/bin/pushover
>>     repeat every 2 cycles
>>
>> At random in the middle of the night, I'd occasionally get an alert 
>> that ping failed. Device was up and responsive. Modified it to:
>>
>>     check host resolver2 address 192.168.1.13
>>     every 5 cycles
>>     if failed ping4 count 5 timeout 5 seconds
>>     then exec /usr/local/bin/pushover
>>     repeat every 2 cycles
>>
>> Which seemed to work for a while, but the failures returned. So 
>> modified it to this:
>>
>>     check host resolver2 address 192.168.1.13
>>     every 5 cycles
>>     if failed ping4 count 5 responsetime 200 milliseconds timeout 10
>>     seconds
>>     then exec /usr/local/bin/pushover
>>     repeat every 2 cycles
>>
>> However, this causes an immediate failure with 'Speed failed'. I 
>> increased responsetime to 1 second, same. Increased to 10 seconds, 
>> same. The specific error logged is
>>
>>     NovĀ  5 09:52:45 ntpviz monit[676]: 'resolver2' response time 6.173
>>     ms doesn't match limit [time = 10 s]
>>
>> I may be misunderstanding the difference - or the interaction - between 
>> responsetime and timeout. The device is just a raspberry pi zero, with 
>> a typical responsetime under 10ms, and never under appreciable load.
>>
>> The manual states
>>
>>     The RESPONSETIME parameter sets the response time limit
>>
>> Is that the response time limit for any given ping, or if for example I 
>> use a count of five, is it the time limit for all five attempts 
>> accumulated? Even if it were the latter, since the device is reachable 
>> in under 10ms, a count of five - assuming 1 second between each ping 
>> sent (is that a correct assumption, or are they sent without pause?) - 
>> would still complete in well under 10 seconds.
>>
>> For reference, here's the results of a 60 second ping from the monit 
>> server to the device:
>>
>>     root@ntpVIZ: /etc/monit/conf.d # ping -c 60 192.168.1.13
>>     PING 192.168.1.13 (192.168.1.13) 56(84) bytes of data.
>>     [elided]
>>     --- 192.168.1.13 ping statistics ---
>>     60 packets transmitted, 60 received, 0% packet loss, time 59094ms
>>     rtt min/avg/max/mdev = 2.342/6.376/13.879/2.335 ms
>>
>> Thanks for any insight.
>> -- 
>> Paul Theodoropoulos
>> www.anastrophe.com
>

-- 
Paul Theodoropoulos
www.anastrophe.com