On Wed, Jun 22, 2005 at 12:56:00PM -0400, Jeffcoat, Al wrote:
Here's where the problem starts. Windows has been generating a lot of red events. The execs overreact to red events that are really not problems. We've tired to explain that just because something is red, doesn't mean that a system is down. The windows admins are tuning it down as best as they can, but it still happens.
So my thought was to have two hobbit's, one that receives client data for our engineers to see and work with, and then another that just does the network tests, and this would be what our execs see. Both would be sitting on the same server, running in separate directories, and separate virtual servers. What I'm trying to get to is two views, an executive view, and a technical view.
This sounds like what pagesets are for. Basically, you can define an alternate layout of the hosts on your Hobbit webpages, and when generating the webpages for the new layout you can remove certain tests that are likely to cause problems - e.g. the "msgs" column from your Windows boxes.
This is described in the bbgen(1) man-page, look for the "BUILDING ALTERNATE PAGESETS" section.
When generating the executive webpages, you can also use some of the bbgen options, like "--ignorecolumns=msgs" if you just want to completely remove any "msgs" column from the view, or the various --noprop* options to keep a non-green status from affecting the color of higher-level pages.
A very simple solution might be to just use the "NK" page as the executive webpage. Statuses appear on the "NK" page only if they are listed in the "NK:testname,..." tag for each host, so you can pick out the critical tests and put them on a single page for your executives. Then they have a one-page overview of all systems - I think that would appeal to the group of people you're targeting. You don't need to run multiple pagesets for that - the NK page is generated by default.
Running two full Hobbit's is overkill - you already have the data, it's just a matter of presenting it.
Regards, Henrik