I know for our environment this won't work. From what I understand (I am not a web dev dude) it is common to run an IIS server with 1 or even multiple IP's but still configure each site via the Host Headers. It is also common to have a load balancer in front of your WebServer Clusters as well. So actually the URL is pointing to a Virtual IP Address hosted on a Load Balancer - then it gets submitted to one of the individual web servers.
With that said...how do I send the HTTP monitor to an IP address where I want to test each individual server so I can be alerted when any single server within the Web Cluster has an issue? Remember each server within the cluster hosts multiple sites and uses host headers in order to get the user to the proper site. So entering in say a private/internal 10.x.x.x address to a specific web server within the cluster - it still needs to send the host header in order to load the proper page that I am looking for.
Hence, if I use only the URL it is going to resolve to the VIP on the load balancer which will then still land on any one of the web servers within the cluster and yet still need a host header in order to get to the proper page.
So, how do I resolve that issue?
Is there any update on this feature request? We make heavy, heavy use of Apache HTTP Server Name-based Virtual Hosting and we absolutely have to be able to specify the HTTP Host: header in order to properly perform a HTTP Monitor test against our web servers.
We can't simply hard-code a Url because that will send us through a load balancer and testing the individual nodes *behind* the load balancer is necessary.
We can't put the IP address of the node in the Url field because that will only test one of possibly dozens of Name-based virtual hosts we have on these servers. We need to test and validate them all.
I am shocked that with the plethora of options available in the HTTP Monitor configuration that HTTP Host: header is not something we can customized.
I tried using the Proxy Address: workaround described above but it didn't work for me - I got "The request was aborted: The operation has timed out."
Indeed, it is pretty poor that Solarwinds, supposedly an enterprise level monitoring solution does not seem to support the basic features of the HTTP protocol. If it does, I can't find the documentation. And if you want to know how silly this is, HTTP isn't a new or obscure protocol, it's part of the building block of the internet for the past 15 years.
We have 500+ sites all terminating on the front end load balancers and they talk back to webfarms using host-headers, this means we only need 1 IP per webserver.
Of course we can monitor using the URL but that only hits the load balancer. We need to check each of the back end servers. Even the dirt cheap host monitor has this option, you should get your product guys to download that because it's 500% more configurable than this 'enterprise' solution.
So right now I have to tell my colleagues that Solarwinds cannot run the most basic of HTTP requests. Is this April Fools or what?