How to monitor two internet connections to detect outages and switch to the active one when one is down

Question

I have two different internet connections on my house (from two different ISPs). And I use a TP-Link ER605 Multi-Wan router to manage both connections and deliver a single internet connection to my wired home network.

This router has a single job: When the main connection goes down, it redirects all the traffic to the backup connection, and I just don't notice any outage at all when it happens.

The thing is: I would like to get some kind of notification when my main connection is not working so I can check if something is wrong on my end, and if this situation persists for some time, I can call my ISP to ask for technical support.

I can use "traceroute" (on Linux) or "tracert" (on Windows) to trace the route between a computer on my network and the Google Public DNS (8.8.8.8) to determine which Internet connection I'm using.

tracert -h 2 -4 -d 8.8.8.8

traceroute -4 -n -m 2 8.8.8.8

If the second hop is 192.168.15.1, it means that my router is using the main connection to get to the Internet, but if the IP address is 192.168.0.1, it means that the main connection is offline and router is redirecting the traffic over the backup connection. (On this specific case, it's on the third hop on the linux example, because it's a VM on another NAT network).

The problem is: I don't know how to automate that verification, and get some kind of notification depending on which connection is in use. It could be an e-mail like "Main connection is down" and "Main connection restored".

I have my Laptop running Windows 10 running 24/7 and I also have a Raspberry Pi (with Raspian Linux) that I could just leave connected all the time for this specific task. So both Windows and Linux solutions would work for me.

Answer 1

Bash Script to Monitor WAN

This script will write to a log file. You can add or substitute a function to integrate sendmail from a command line. I would recommend that as an additional separate forum question. You will likely want to have some number of lines from the logs to attach to any email you ultimately send yourself.

1. Script will run continuously until killed
2. Script writes to a log file only when the connection changes (UP/DOWN).
3. At startup if both connections are down that will be logged.
4. Log file is created in the same directory as the script.
5. Log file follows script name but with a .log extension.
6. New log file entries are written above old ones.
7. Script will log when a connection comes UP.
8. Script will log if that connection goes DOWN.
9. Script will check a connection is still UP every x seconds (currently 5).
10. If both connections are DOWN script will check for UP every x seconds (currently 5).
11. Remember when creating the bash script to make it executable (chmod +x file.sh)
12. You can uncomment line 47 if you want to simulate your connection dropping after x successful tests (currently 5)
13. To maintain log entry column alignment, when populating 'isp[]' and 'state[]' pad shorter strings so the values in the same array have the same character lengths (e.g. 'Vivo ' 'Claro' instead of 'Vivo' 'Claro').
14. If you need to debug to console, add the option '-x' at the end of the first line ('#!/bin/bash -x')
15. Lots of other posts were adapted for this script, I'll need to reference some of them later.

#!/bin/bash

host='8.8.8.8'
wanip=('192.168.15.1' '192.168.0.1')
isp=('Vivo ' 'Claro')
state=('UP  ' 'DOWN')

scriptpath="$(realpath "${BASH_SOURCE[-1]}")"
scriptdir="$(dirname "$scriptpath")"
logfile="$scriptdir/$(basename "${scriptpath##*/}" .sh).log"
len="${#wanip[@]}"
touch "$logfile"

function addtolog () {
  dates="$(date '+%s')"
  datel="$(date -d@"$dates" '+%Y-%m-%d %T')"
  logentry="$dates: $datel  ${isp[$1]} ${state[$2]} ${wanip[$1]}"

  echo "$logentry" | cat - "$logfile" > "$dates" \
    && mv -f "$dates" "$logfile"
}

function tracert () {
  arr=("$@")
  for i in "${arr[@]}"
  do
    traceroute -n "$host" | \
      grep -F -f <(printf "%s\n" "$i") \
      >/dev/null 2>&1
    (( "$?" == 0 )) && return 0;
}

if ! (tracert "${wanip[@]}"); then
  for (( i=0; i<len; i++ ))
  do
    addtolog "${i}" 1
  done
fi

while :
do
  if (tracert "${wanip[@]}"); then
    for (( i=0; i<len; i++ ))
    do
      c='1'
      while (tracert "${wanip[i]}")
      do
        (( c < 2 )) && addtolog "${i}" 0
        c=$(( c + 1 ))
        sleep 5
#        (( c > 4 )) && echo "Simulating Dropped Connection" && break
      done

      addtolog "${i}" 1
    done

  else sleep 5; fi
done

NB

While I tested the code and tried to focus on your use case, please note that I am not, by any stretch, qualified to be writing code for anyone, am not properly trained, and have likely broken more best practices than I know while coding this.