How are packet loss and RTA Calculated?
Posted: Tue Dec 04, 2012 1:34 am
Hi all,
Trying to figure out how the host check packet loss/rta figures are calculated. I'd also like to know how the ping service check packet loss figure is calculated. I'm trying to tune our Nagios instance and, it appears to me at least, that the default host-based (check-host-alive) packet loss figure of Warning=80% seems surprisingly high. I know this is configurable, but there must be a reason why this is the default.
With regard to the ping service check (check_ping), the values are closer to what might seem reasonable (warning = 20% loss, etc). However, I'm getting Warning alerts in my config saying that I have a packet loss value of either 16%, 28%, 37% or 44%. All of these are actual examples. As far as I can tell the ping check sends 5 "pings". I can't reconcile how the packet loss percentages are arrived at, given that 1 packet lost should equal 20%, right?
Happy to be directed to documentation if that answers my questions, but I haven't been able to find any so far. All help appreciated.
Many thanks,
Ben.
Trying to figure out how the host check packet loss/rta figures are calculated. I'd also like to know how the ping service check packet loss figure is calculated. I'm trying to tune our Nagios instance and, it appears to me at least, that the default host-based (check-host-alive) packet loss figure of Warning=80% seems surprisingly high. I know this is configurable, but there must be a reason why this is the default.
With regard to the ping service check (check_ping), the values are closer to what might seem reasonable (warning = 20% loss, etc). However, I'm getting Warning alerts in my config saying that I have a packet loss value of either 16%, 28%, 37% or 44%. All of these are actual examples. As far as I can tell the ping check sends 5 "pings". I can't reconcile how the packet loss percentages are arrived at, given that 1 packet lost should equal 20%, right?
Happy to be directed to documentation if that answers my questions, but I haven't been able to find any so far. All help appreciated.
Many thanks,
Ben.