Ah, the memories of AWS outages in the early days. /s

Sorry for them. I assume usage spiked up (again), and of course it's not exactly easy to handle particularly aggressive spikes.