Quote Originally Posted by Yogol View Post
And companies that plan ahead, have the possibility to add extra servers in times of need for a minimum cost, cause they rent their servers at some 3rd parties server-farm (my company uses IBM, but there are many, many others) so they can request alot of extra servers just for days like these.
Calling it "minimum cost" is more than a little disingenuous.

The software for whatever you need those failover servers to do has to be installed on the servers ahead of time....you don't just automagically hook up those servers and get automagic extra capacity. That costs money, as the 3rd party won't let your stuff just sit on their servers for free. Those additional servers also have to be worked into your failover infrastructure (that is, the routing to them, allocation of which servers take over which functions for which primaries, load-balancing among all of the above, etc). That process is also not free.

COULD all of those things have been done? Of course. Was the cost to do so acceptable, given the financial risk involved in you waiting an extra few hours to patch? I dunno, and neither do you...neither of us writes contracts for Turbine, neither of us can say what this is realistically likely to cost them in lost revenue.

I'll say it again: 100% flawless implementation and redundancy, resulting in 100% 24/7/365 stability, even during major updates, is a constantly-moving target that NOBODY ever truly hits. Anybody who tells you differently is probably a vendor trying to sell you something.