Re: [SAGE] Color identifiers and products like Nagios ...

["Ian C. Blenke" <ian@blenke.com>]

Jennifer Davis wrote:

> So reading about the planned changes with Firefox, Opera, Konqueror, 
> and IE7 (reference 
> http://blogs.msdn.com/ie/archive/2005/11/21/495507.aspx).. especially 
> the idea of color coded address bar to report phishing and other 
> possible negative aspects to a potential website, and then reading 
> some of the comments about color-blindness, and poor color choices, I 
> wondered about the use in monitoring tools like nagios.


Thunderbird also has a neat feature where you can colorize entries by 
adding X-Keywords headers of $Label1 through $Label5 in your .procmailrc 
recipies. This makes for a great visual hint when scanning your mailboxes.

:0 fhw
* ^From:.*sombodyimportant@mycompany.com
| formail -I "X-Keywords: \$Label1"

> Is the standard "red" - critical state, "yellow" - warning state, and 
> "green" - good/normal state good indicators for monitoring.  I 
> understand that there is an associated value with those colors, but is 
> that American-centric?


This is really a human interface design issue. Most folks in the US have 
been trained with those three colors straight out of driving school. 
Most children learn early on that a yellow light means "go faster".

That does leave out those with color blindness or other maladies that 
make it difficult to discern between such coded alert levels. Ideally, 
you would use a background image pattern, icon, or animation along with 
the color to denote a function.

What of animated icons or other interface design issues? There is an 
entire discipline of Human Interface Design that should be applied to 
such things. As a system administrator, I'm really only concerned with 
my target audience and what they can grok.

> If you were designing a brand new, from scratch tool/application for 
> system administrators .. would you follow the standard red, yellow, 
> green rules?  If not, why not?


I would hire a designer with some understanding of human interface 
guidelines to design as simple an interface as I could to appease my users.

Simple. Clean. Efficient. Without getting in the way. Apple might be a 
prime example, though they often break that last rule.

> In part, I'm wondering about the information one would want to convey 
> to the rest of the people in an organization outside of the IT 
> department, who don't necessarily want to understand the technical 
> aspects but want to understand how/if they are affected.  (not wanting 
> people to attribute their possible experience with an announced 
> problem, but also not wanting to have to repeat multiple times with 
> updates to different people)


What you're really suggesting is a general status page with a Homeland 
Security style colorized alert assessment might give users an idea of 
the availability of common services they use. From "Your mail service is 
mildly impacted and may be a bit slow" to "Your mail service is horribly 
impacted and may not be back up until tomorrow."

To this end, you're really acting as any other service provider. If your 
cable TV goes out, you can call it in and open a ticket or go to the 
technically advanced cable provider's website and see what the status is 
and who is affected.

I'm kind of keen on the idea of a "plaid" alert level myself.

- Ian C. Blenke <ian@blenke.com> http://ian.blenke.com/