If you were hoping to spend the last couple days catching up with your favorite websites while maybe streaming a Netflix movie, you’ve probably discovered by now that you had pretty lousy timing for a lazy web-surfing weekend.
On Friday, bad storms in the Southeastern United States knocked out a big swath of Amazon Web Services hosting infrastructure that led to full and partial outages for the likes of Instagram, Pinterest and Netflix. And just as the web was starting to bounce back from that, a planned “leap second” — the addition of one second onto the world’s official clocks to keep us on track with the actual rotation of the Earth — tripped up the functions of a number of other major websites including Reddit and Gawker. Wired’s Cade Metz has written a good report of what happened there.
To me, the real takeaway here is how fragile so many of today’s most popular and relied-upon web services actually are. Of course, with so many millions of people on the web, and so relatively few people with the ability to write code, it’s kind of amazing that it all generally works as well as it does. But this past weekend’s meltdowns show how much work is left to be done.