There is a special, though common, case when wanting to monitor multiple websites that are hosted on the same server, under the same IP. This is called virtual hosting.

http://en.wikipedia.org/wiki/Virtual_hosting Virtual hosting is a method that servers such as web servers use to host more than one domain name on the same computer, sometimes on the same IP address.

Zenoss cannot have >1 device with the same IP. You cannot create a separate device for each website in this instance, because you cannot have >1 devices resolve to the same IP in Zenoss.

You may be trying to add a second, third, etc device that had the same IP as your base server.

Add multiple httpmonitor datasources in a single local template. The datasources will be inherited by the device that is hosting the many websites.

The following are more detailed instructions, since it might be hard to understand the above description:

1. set up the entire server as a normal device being monitored via SNMP, WMI, or SSH (likely, you already have this)
2. create the device with any device class really, a /devices/server/* device class will usually do
3. in this example, the web server device will be named “atlas.loc”

1. while navigated to the web server (atlas.loc) device, add a template
2. this will come out to be a ‘local copy’ template, which means it is not applied to an entire device class, but just a single device
3. only one template is needed, called “webPageChecker” (or whatever you wish)

1. name each datasource after the various websites that you will monitor
   1. Ex: worldatlas.com, globeconf.com, etc (these are server off same server)
2. remember to choose HttpMonitor type on each datasource
3. Host Name: worldatlas.com
4. IP Address: ${dev/manageIp}
5. Cycle Time (seconds): sets how long it will wait between checking the website (default 300);
   1. this effects zenoss performance; dont make it too low or your free CPU and UI responsiveness may drop
6. Url: /corp/nctrip/main.html (or whatever page)

1. you are now able to put all websites response times on a single graph, if wanted!

From ticket #6616