Most of the servers we dealt with were moved at "silly hours" but, since we have customers from all around the globe under various different brands, it's impossible to please everyone. We chose to move the 'Kylie' and 'PHP5' (the latter being phased out shortly as PHP5 is now standard) servers earlier in the day as the impact would be lower as these servers only host in the region of 500 clients each compared to 1500+ on our larger Quad Xeon machines.
As for "always having downtime", you certainly shouldn't be. I'm sure our other users who are resident on this forum will concur. If you're having problems, please email with as much info as possible

.
I should also add, in our defense, that this was planned downtime required to ensure and improve the ongoing stability of our service. Many shared hosting providers in the same price bracket will have hours of "unplanned" downtime every month. The grass is not always greener.
In addition, myself and Darren (both full time employees, not part time uni students, IT dept drones, etc) now live within walking distance of the datacentre (our own private suite, custom specced in a top spec carrier neutral facility). How many virtual hosting providers can claim that, with their servers in Manchester, Florida or Texas?
I hope that puts your minds at rest. I do apologise once again for any inconvenience caused. If you have any queries, please raise a ticket.