You never quite know exactly how something will work until you actually start using it. This is quite true with regard to a website that the world uses.
When we bought Mira (the webserver here at HQ) we specifically overpowered her, and I knew that. Seeing the prices for cloud instances at Amazon we went with what is called a "Large" instance which, really was about half of Mira's size.
Over the last 24 hours we've experienced a CPU utilization that averaged around 75-80%, maxing out at 100% for a few hours at a time. Some of you reported things like uploading data to the database taking 10-15 minutes, but finally working. Others reported things such as not being able to get a DSS image from VSP, but when Matt & I tried it this morning, we were able to get such an image.
After some research and a good bit of discussion with Arne, Matt, Rebecca, Will, and Mike Simonsen, everyone seemed to agree that we needed to at least try to upgrade the cloud server instance. Fortunately Amazon makes this relatively painless.
So, at 2:00PM, Eastern Standard, 18:00hrs UTC, today, we will take the website down for about a half hour and upgrade the instance to see what kind of difference this makes in performance.
We have heard you folks with regard to performance problems you've experienced over the last 24 hours. These problems are different and separate from specific things that need to be fixed (like getting into the chatroom from the website page, as a good example). We are taking immediate steps to answer these performance concerns.
We thank you for your patience and support as we try to make things better!
AAVSO Astronomical Technologist