We’ve spent a lot of time since we launched last Spring improving our software load-handling. To give you an example of how our infrastructure has improved, here are the CPU levels on our primary database server during the iPhone OS 3.0 event today. This liveblog alone was sustaining over a thousand people at a time (remember, those are real-time numbers of visitors, not totals). As you can see, our database rarely got over 2% CPU for all the events running simultaneously this afternoon.
There is lots more we can do to further cache our sites, but for now, I’m happy to report that we are looking very stable under those spiky loads 
