Amazon’s cloud has become pretty notorious for having outages, particularly in its pesky East Coast data centre. It happened again last week, taking down sites like Quora, Pinterest, HipChat, Heroku and others.
Research obtained by Business Insider shows that 40% of Amazon’s biggest cloud users are not ready for the next outage—and it’s pretty much their own fault.
We asked Israeli startup Newvem to come up with some insight about Amazon’s last outage. Newvem is beta testing a service that helps companies plan big AWS sites (how many instances of servers they really need, stuff like that).
Newvem poked around and discovered that 40 per cent of Amazon’s biggest Web customers made the outage worse on themselves because they don’t follow rule No. 1 in computing: make backups.
And they aren’t using free or low-cost services that Amazon already provides to do so. So if Amazon falters, they can’t move their site elsewhere and keep it plugging along.
“These figures really shocked us, because Amazon provides many tools (free and paid) that help protect and even keep a user’s cloud running during an outage,” says Cameron Peron, Newvem’s VP of business development.
Amazon offers tools like Elastic Load Balancers, which automatically shift traffic around, and Snapshots, which automatically make backups. But they don’t work if websites don’t use them.
Just to give you an idea of how insane that is, Amazon charges peanuts for these services, a mere $0.025 per Elastic Load Balancer-hour or $0.008 per GB of data processed by an Elastic Load Balancer. That’s a fraction of a penny. Snapshots for a medium-sized site—say, 100 gigabytes of data—might run a whopping $36/month, Amazon says.