System and throughput sizing are as much an art as a skill. An old rule of thumb stated you needed to size a system at 200% of the average throughput. The feeling was this would accommodate upwards of 90% of any surges. Generally speaking occurrences of greater volumes would be so few an and far between that it would not be cost effective to size any bigger. The exception to this would be if any disruption to your throughput would be catastrophic to your business, you would size to 99% availability level if you could afford the cost. There was always a lot thought and sweat that went into these decisions. It is virtually impossible to have 100% availability of any resource, but with proper evaluation of the relevant factors, you can come close.
Cheers

Difficult to communicate when no-one can get through to your website.

Well, this download problem will solve itself, if Krembil jut continue to wait, doing nothing. The more people that leave the project, the easier it will be to get work.

That's a solution of course, although not a very good one.

Ah well, getting tasks is no big problem. The big problem is getting them downloaded. Downloading the tasks, is like swimming up the Niagara Falls.

this is getting fun to watch unfold.

getting a free autoclicker to press "retry now' every 60 seconds solves the issue for my PCs but not for mobile...