Thanks for your reaction, but I replaced my initial message with this one because I didn't know how to completely delete the thread.
I was suddenly getting very bad ping-results to one modem I just upgraded.
I am doing this from Zabbix and I can reliably sense packetloss this way.
This was the first time I had a false positive in years....
The reason, however, wasn't the upgrade itself but the fact that I pinged that router from the console of that same monitoring server to see when it came online...
I just forgot about that ping (I sometimes have about 30 windows open in putty) and it went on for hours....
AVM probably has some limiter on it.
I normally get reliable results.
When I wrote that "packetloss" script I knew I shouldn't stress the other endpoint so it's a carefull burst of 5 pings with an interval of a 300 milliseconds
Zabbix calls this script each 20 seconds for many connections...
# cat `which packetloss`
Code:
#!/bin/bash
PACKETLOSS=`ping -q -c5 -i0.3 -w4 ${1} | grep transmitted | grep -Eo '[0-9%]* packet loss' | tr -cd '0-9'`
[ -z "${PACKETLOSS}" ] && PACKETLOSS=100
echo "${PACKETLOSS}"