Page 1 of 1

website monitor wizard failure

Posted: Fri Apr 22, 2011 4:57 pm
by jrandall
I use tomcat to provide web services for an application. I specified a non-standard port number (29772) since I have multiple instances of this application running on the same server. I used the monitoring wizard in Nagios XI to set up a monitor for the web page content of this instance of the application . The web page content portion of the monitor works fine. However, there is another portion of the monitor that appears to perform a check of the host on the default http port. Since I don't run IIS, apache or any other complete web server on the box, this check fails.

I do not see any way from either the wizard or from the Core config manager to change the default http port.

I was able to fake out this host check by changing the IP address of the host in the host statu detail page. Changing the IP address of the host did not affect the web content monitor. This is not a viable long term solution.

How do I re-direct the default http port to my alternate port? If that can't be done, how do I remove this host check but keep the web content monitor?

Thanks,
John

Re: website monitor wizard failure

Posted: Mon Apr 25, 2011 4:01 pm
by rdedon
One of our team has been working with Tomcat recently and I will see if he has any insight into this.

Re: website monitor wizard failure

Posted: Tue Apr 26, 2011 4:20 pm
by tonyyarusso
The host check is probably using check_xi_host_http, which in turn uses the check_http plugin. If 29772 returns an HTTP response, then add the -p switch as an argument, as noted in check_http's help output. If not, change the check command for that host in the Core Config Manager to a generic TCP check against 29772, or whatever else you want to do.

Re: website monitor wizard failure

Posted: Wed Apr 27, 2011 5:17 pm
by niebais
tonyyarusso wrote:The host check is probably using check_xi_host_http, which in turn uses the check_http plugin. If 29772 returns an HTTP response, then add the -p switch as an argument, as noted in check_http's help output. If not, change the check command for that host in the Core Config Manager to a generic TCP check against 29772, or whatever else you want to do.
I usually modify my options to look like this:

check_http -H hostname -u /index.html -f follow -w 25 -c 30. If it's an ip just change it to check_http -I ipaddress. Sometimes you have to modify the options after the XI wizard sets it up.

Re: website monitor wizard failure

Posted: Thu Apr 28, 2011 10:29 am
by admin
Thanks for the tip!