False reporting on bash script - linux

Question

Context
I have a script that pulls down data from a web server using wget, checks the filesize of the index.html page using ls -la, if it is a certain size using grep -o, it is assumed the page is non-functional to customer requirements.
Code
BADSERVER=()
FILE=SERVER
echo '' > $FILE.txt
<pull a list of hosts down>
if ! ssh -n -f $device -i <sshkey> "yes | rm index.* ; wget localhost:8080 ; ls -la index.html | grep -o 12282" > /tmp/$FILE 2> /dev/null; then
echo "Unable to connect to $device"
BADSERVER+=($device)
echo '' > /tmp/$FILE
elif cat /tmp/$FILE | grep '12282' ; then
echo 'Page is non functional to customer requirements on $device'
BADSERVER+=($device)
echo '' > /tmp/$device
else
echo 'all is fine on $ip'
echo '' > /tmp/$FILE
fi
done
Problem
The script reports that some webservers have the filesize, when manually connecting I can see that the file is not the same size.
Script traceback
+ ssh -n -f webserver.1014 -i <omitted> 'yes | rm index.* ; wget localhost:8080 ; ls -la index.html | grep -o 12282'
+ grep 12282
+ cat /tmp/SERVER.txt
+ echo 'All OK on webserver.1014'
+ echo ''
+ cat /tmp/SERVER.txt
+ cat /tmp/SERVER.txt
+ ssh -n -f webserver.1015 -i <omitted> 'yes | rm index.* ; wget localhost:8080 ; ls -la index.html | grep -o 12282'
+ cat /tmp/SERVER.txt
+ grep 12282
12282
+ echo 'Page is non functional to customer requirements on $device'
Page is non functional to customer requirements on webserver.1015
+ BADSERVER+=($device)
+ echo ''
+ cat /tmp/SERVER.txt
+ cat /tmp/SERVER.txt
+ ssh -n -f webserver.1016 -i <omitted> 'yes | rm index.* ; wget localhost:8080 ; ls -la index.html | grep -o 12282'
+ cat /tmp/SERVER.txt
+ grep 12282
+ echo 'All OK on webserver.1017'
+ echo ''
+ cat /tmp/SERVER.txt
+ cat /tmp/SERVER.txt
Actual content of webserver.1015
-r--r--r--. 1 --- --- 25799 Mar 30 23:38 index.html

Answer 1

Is the server up & running at webserver.1015 ?
Why are you using ; instead of && ?
If previous command failed with ; the second one will run.
But with && the second one will not run.
Don't parse the output of ls command.
Extracting only file size with wget
wget http://localhost:8080/index.html --spider --server-response -O - 2>&1 | sed -ne '/Content-Length/{s/.*: //;p}'
Or with curl
size=$(curl -sI "http://localhost:8080/index.html" | awk '/Content-Length/{gsub("\\r", ""); print $2}')