ISP connectivity logger

[five.five-six]

I have a customer who is having lots of, what I suspect are, ISP connectivity outage problems. What I have experianced while on site are short service interruptions. They are causing a lot of problems for the customer and consequently me.

I want a stand alone solution, something I can setup and leave in premise for a week or two, or possibly remotely continuously ping their gateway. IDK, but I suspect they are having clusters of short service interruptions and I need to verify or rule out my suspicions.

Any ideas?

[command_liner]

How about a headless NUC running up-to-date Ubuntu 16.04 LTS. Make a user, and have that user run a cron job. Once a minute ping ten packets to someplace like google.com, amazon.com. Record the times for the packets.

Graph the result. Look for obvious errors.

[Ninety]

Who is the isp? Sounds like a bad switch on their end

Sent from my HTC Desire 626s using Tapatalk

[SkyHawk]

Pingplotter pro, or a simple dos script. Pingplotter is nice because you get graphs and can set a very small sample interval. You can set multiple targets, and I would ping the customer edge device (firewall or router), the ISP gateway, and something beyond like the Google DNS server (8.8.8.8). Use tracert to get the IPs that you need.

Pingplotter Pro version is free for two weeks. Otherwise the perpetual free version only works against one target.

A dos script is simple. Run several CMD windows and in each, ping the same targets every 10 seconds. Run a find against the output, and if you get 'timed out' then log the time and a note to a log file.

https://www.pingplotter.com/download

[L4D]

Probably uverse

[SkyHawk]

Here is a DOS batch file you can use. It will ping 4 samples every 10 seconds (you can change this) and if ALL SAMPLES FAIL (100% loss) from one of the checks, it will log the fact to a file with time and date stamp and continue looping.

You need 'sleep.exe' from a MS Windows resource kit, otherwise use something else to waste 10 seconds between loops. If you need sleep.exe, PM me.

This script with the basic settings is good if you suspect you are frequently going down hard for at least a couple of seconds (or 10 - 20+ seconds). If you suspect you are flapping up/down at shorter intervals and want to capture that, then lower the loop interval to 1 second and only send 1 ping packet, or lower it as appropriate. It will only log ping failures with 100% loss.

If you want to run this against multiple targets, make a copy of this file for each target with the appropriate IP, and run them in separate CMD windows. You can run them all from the same directory and they will log to the same file, or you can run them in separate directories so they each have their own log file, or you can mod the script so each target had it's own log file but all in the same directory.

@echo off
cls


::***************************************************
::*SETUP SCRIPT VARIABLES
::***************************************************

::ISP UPSTREAM ROUTER IP TO CHECK
set pingip=192.168.0.61

::SLEEP TIME IN SECONDS BETWEEN LOOPS
set sleeptm=10

::NUMBER OF PINGS PER TEST
set npings=4

:: !!-DO NOT CHANGE THIS SETTING-!!
set loss="100%% loss"

::***************************************************



::UPDATE LOG
echo %DATE%  %TIME%  -  Monitoring started for IP %pingip% >> ping.log


::(RE)START
:start

::PREPARE SCREEN
cls
echo %DATE%  %TIME%  -  Pinger (re)started...


::*************************************************
::MAIN MONITORING LOOP
:loop

sleep %sleeptm%

cls
echo SENDING PINGS TO IP %pingip%
::PING THE ISP UPSTREAM ROUTER
ping -n %npings% -w 100 %pingip% | find %loss% > nul


::IF THE PING IS BAD:
::GOTO THE FAILED LINK VERIFICATION SEQUENCE
if not errorlevel 1 goto failed

echo PING WAS GOOD
goto loop



:failed
echo PING FAILED TO IP %pingip%
echo %DATE%  %TIME%  -  PING FAILED TO IP %pingip% >> ping.log

goto loop

[five.five-six]

I like that, I have a stick PC running windows I could run that on. Would sleep.exe run on a raspberry like device?

Here is a DOS batch file you can use. It will ping 4 samples every 10 seconds (you can change this) and if ALL SAMPLES FAIL (100% loss) from one of the checks, it will log the fact to a file with time and date stamp and continue looping.

You need 'sleep.exe' from a MS Windows resource kit, otherwise use something else to waste 10 seconds between loops. If you need sleep.exe, PM me.

This script with the basic settings is good if you suspect you are frequently going down hard for at least a couple of seconds (or 10 - 20+ seconds). If you suspect you are flapping up/down at shorter intervals and want to capture that, then lower the loop interval to 1 second and only send 1 ping packet, or lower it as appropriate. It will only log ping failures with 100% loss.

If you want to run this against multiple targets, make a copy of this file for each target with the appropriate IP, and run them in separate CMD windows. You can run them all from the same directory and they will log to the same file, or you can run them in separate directories so they each have their own log file, or you can mod the script so each target had it's own log file but all in the same directory.

@echo off
cls


::***************************************************
::*SETUP SCRIPT VARIABLES
::***************************************************

::ISP UPSTREAM ROUTER IP TO CHECK
set pingip=192.168.0.61

::SLEEP TIME IN SECONDS BETWEEN LOOPS
set sleeptm=10

::NUMBER OF PINGS PER TEST
set npings=4

:: !!-DO NOT CHANGE THIS SETTING-!!
set loss="100%% loss"

::***************************************************



::UPDATE LOG
echo %DATE%  %TIME%  -  Monitoring started for IP %pingip% >> ping.log


::(RE)START
:start

::PREPARE SCREEN
cls
echo %DATE%  %TIME%  -  Pinger (re)started...


::*************************************************
::MAIN MONITORING LOOP
:loop

sleep %sleeptm%

cls
echo SENDING PINGS TO IP %pingip%
::PING THE ISP UPSTREAM ROUTER
ping -n %npings% -w 100 %pingip% | find %loss% > nul


::IF THE PING IS BAD:
::GOTO THE FAILED LINK VERIFICATION SEQUENCE
if not errorlevel 1 goto failed

echo PING WAS GOOD
goto loop



:failed
echo PING FAILED TO IP %pingip%
echo %DATE%  %TIME%  -  PING FAILED TO IP %pingip% >> ping.log

goto loop