Kobe Bryant died in the crash of a helicopter he chartered, this morning. TMZ.com broke the news and the load brought down the TMZ servers... they couldn't respond to the traffic. And now, hours after his death, his Wikipedia page is highly unresponsive as people visit and edit it.
I got my first taste of managing extreme server load when I worked at Apple, after a Steve Jobs Keynote speech. But we always anticipated that load. When Michael Jackson died, in 2009, I was working with my Adjix startup and even simple URL redirects overloaded my servers.
When a celebrity unexpectedly dies, server load goes through the roof. I call it Celebrity Server Overload. You can plan for it to happen, you just don't know when it will happen. But you can scale up as the traffic grows.
The other end of the spectrum is when a business is brand new – they have no experience on scaling up. Facebook could manage their growth, in the early days, because they brought on new users one college campus at a time. But even smart developers can't handle a high load, off the bat, without any past data points. We saw this with Twitter's fail whale and the launch of Obamacare. Semper paratus, or as we used to say in the Boy Scouts, Be Prepared.
Kobe Bryant: August 23, 1978 – January 26, 2020.
No comments:
Post a Comment