[Nagiosplug-devel] check_http feature for web cluster

[Garrett Honeycutt]

You should probably define your load balancer as a host and check it
using check_http -H www.domain.name and have each machine in your
cluster setup as a host with a service description that runs check_http.
This is what I do for my web cluster. I think that extended the check to
check multiple hosts is not really the correct way to go about things.

-g


On Mon, 2005-10-03 at 13:20 +0300, Aurimas Mikalauskas wrote:
> Hello,
> 
> I'm missing one feature in check_http nagios plugin that would help
> monitoring different virtual hosts on clustered web servers. The
> situtation is simple:
> 
> I have two virtual hosts www.first-host.com and www.second-host.com
> both round-robin resolving to 10.0.0.1, 10.0.0.2 and 10.0.0.3.
> 
> I want to check each virtualhost on each cluster. If I do
>   ./check_http -H www.first-host.com
> 
> then it will likely round-robin each check to a differend cluster and
> most likely at least one of them will be alive and return correct
> data.
> 
> If I do "./check_http -I 10.0.0.1" then I will be able to check only
> the first virtual host.
> 
> If you added one more variable that is the virtual host (even if I use
> -I 10.0.0.1) somewhere near 824 line after the "Host: ", the plugin
> would be much closer to perfetct. On the other hand, you could simply
> expand the <uri> variable so it could be used as
> "http://virtualhost/uri".
> 
> Thank you in advance for your good work!
> 
> Respectfully,
> Aurimas Mikalauskas
-- 
// Garrett Honeycutt
// Sr. Systems Administrator
// 3GUpload.com, Inc.
// 317.472.4969 Office