On Tue, May 01, 2007 at 12:09:26PM -0700, Francisco Valladolid Hdez. wrote:
> Hi folks.
> 
> Reading the Hubert'f Blog, I found this nice post:
> 
> http://www.feyrer.de/NetBSD/bx/blosxom.cgi/nb_20070501_0319.html
> 
> My question go in the next sense:
> 
> How can maintain a high uptime, when severals patches
> are enabled in a time period or maybe in a stable
> version ?

The problem here is that they can only effectively monitor site
availability, not really system uptime.  Theoretically, you could reboot
the system and have it back up before they do their next check.

Also, your systems could be behind a load-balancer of some sort.  Say
you have 3 boxes serving up your website and a load balancing system
fronting the website.  If you upgrade your systems one at a time,
allowing the load-balancer to do it's job properly (not send requests to
the downed system), they would never see the downtime.

Without more information, their stats are only really useful to show
what is popular.  There are too many variables to draw too many
conclusions from the rest of the data.

-- 
Michael Parson
mparson@bl.org