Nagios Monitoring Text on a Website

Question

I was wondering if I could get some help with monitoring text on a website. I.E. If I wanted to monitor google.com for the text "Privacy", I thought I would use the following command:

check_http -H google.com -u http://www.google.com -s "Privacy"

But it is not working. I get "OK" no matter what I put in quotes. I am obviously using either the wrong command or wrong option. Please Help.

Answer 1

Try leaving put the -u. -u gives the path (page) to retrieve, default is "/", it doesn't take the entire URL. Here is my output when I leave it out.

$ ./check_http -H www.google.com -s "Privacy"
HTTP OK HTTP/1.0 200 OK - 0.041 second response time |time=0.040579s;;;0.000000 size=5257B;;;0
$ ./check_http -H www.google.com -s "Privacyblahdibla"
HTTP CRITICAL - string not found|time=0.048169s;;;0.000000 size=5257B;;;0

If you want to get a specific page, use the -u like this

$ ./check_http -H www.google.com -u "/ig" -s "Privacy"
HTTP OK HTTP/1.0 200 OK - 0.166 second response time |time=0.165896s;;;0.000000 size=87843B;;;0

Answer 2

Another option is to use the check_curl from monitoringexchange.org. In reality this is a titch too complicated for what you are trying to do, but I have found it extends the functinality when you need it for parsing data from websites and inserting input

Contents of my customized non-variable check_curl below:

 #!/bin/bash
 PROG=/usr/local/bin/curl
 FILE=/tmp/check_curl
 HALT=PRIVACY

 $PROG -k -s http://www.google.com > $FILE

 STATUS=`grep Error $FILE | awk '{ print $0 }'`

 grep -w $HALT $FILE > /dev/null
 if [ $? -eq 0 ]
    then
      WORKING=`grep Privacy $FILE | awk '{ print $5 }'`
      echo "Works, Returns data of $WORKING"
      rm $FILE
 else echo "Doesn't return $WORKING"| $STATUS"
      exit 2
  fi

Answer 3

We wrote a custom perl script using LWP and HTML::Tree to search for particular strings. We also just md5sum some pages where the content doesn't change.

Answer 4

If I get a problem like this with the check_http plugin, I usually wget the URL I'm looking for the text in and then examine the output.

If you want something that does a bit more than check_http, such as logging into a web site or checking more than one string, have a look at WebInject - it's a nice plugin.

Answer 5

You need to specify the -e option for the -s option to have the desired effect:

$ ./check_http -H "www.eurocarparts.com" -u / -s "no such string"
OK - HTTP/1.1 302 Found - 0.476 second response time |time=0.475776s;;;0.000000 size=17701B;;;0

$ ./check_http -H "www.eurocarparts.com" -u / -s "no such string" -e 302
CRITICAL - string not found|time=1.633004s;;;0.000000 size=17701B;;;0

$ ./check_http -H "www.eurocarparts.com" -u / -s "0870" -e 302
HTTP OK HTTP/1.1 302 Found - 4.032 second response time |time=4.032120s;;;0.000000 size=17701B;;;0

Answer 6

Here are a couple of checks I use.

The first looks for the text CCServerService and reports an error if it doesn't see it:

define command {
    command_name  check_cc_v5
    command_line  /usr/local/nagios/libexec/check_http -H $HOSTADDRESS$ -p 8080 -u /CCServer/CCServer --regex=CCServerService -w 3 -c 5 -t 10
}

The second checks for the text error and reports an error if it sees it:

define command {
    command_name  check_wb_pps
    command_line  /usr/local/nagios/libexec/check_http -S -H $HOSTADDRESS$ -u /onlinestore/booking.php --regex=error --invert-regex -w 30 -c 60 -t 90
}