Gateway Monitoring

[va176thunderbolt]

franco - a gateway monitoring solution that is dependable and reliable is what is needed. I don't believe that ICMP pings are adequate in solely determining that a given gateway is "up" (usable) or "down" (unusable). I've experienced too many instances where apinger reports high packet loss and/or high response time to a gateway while a computer behind the firewall running Pingplotter logs vastly different data.

Maybe we could:
1) wget a (or multiple) user specified url(s) and measure the response time in retrieving the url(s) (say for example, pull the status page my modem and https://www.google.com)
2) open a tcp connection or complete a ssl handshake to a remote host and measure the response time
3) leverage an application like DNS for connectivity & response time (push a query out a connection and ensure it resolves within a user defined time).
4) integrate with vnstat to say push traffic away from a gateway that's capped and reaching the monthly cap.
5) allow me to define a set to tests per WAN connection in determining usability (for example, WAN1 must be able to retrieve https://www.google.com in under 800ms and resolve www.netflix.com via my ISP's DNS server in under 70ms. For WAN2, retrieve the modem's status page in under 10ms and monthly volume under 250GB).

[franco]

I just found this old thread an will revive it for the purpose of advancing the state of apinger:

https://github.com/opnsense/apinger

I'm currently cleaning up the code base, fixing minor and potential problems. The next big hunk is to make it behave when NTP is running, apinger currently does not like that, which is probably most of the "weird" issues that have been spotted.


Cheers,
Franco

[Perun]

Hi

I have still issues with apinger on 18.1.6. I ping my monitoring ip's (DNS server of my isp's) from console and get 15-30ms rtt. With apinger I see sometimes >10000ms at the same time.

How does a cronjob to 'fix' it look like?

Greetz

[prayuth01]

I still do not have these problems. ufabet

thank you

[Perun]

can someone say me how to restart apinger with a command?

[muchacha_grande]

HI,
   I have experienced some issues with apinger in the past, and what I could get as conclusion is that it needs some guaranty on that the PINGs traffic will leave the FW with priority, so even in situations of high upload traffic, the ICMP packets would reach the destination.
   A customer of mine called his ISP telling that the service was faulty and they response that he was using all the upload bandwith. The solution at that moment was to limit the bandwidth consumed by a dropbox sync task on a PC.
   I think that to make sure apinger or dpinger get the best meassurement possible they need to have some QoS in the FW, so that the meassurements results will be more reliable.

Cheers