Web startups crumble under Amazon S3 outage
With what shall I Twitter my daily minutia!?
Amazon Web Services was struck by a temporary outage today, dragging the thousands of web sites which rely on its hosted storage down with it.
Reports of content stored on Amazon's Simple Storage Service (S3) being unavailable or performing poorly began appearing around 5AM PST.
Hardest hit by the outage was the crowd of "web 2.0" startups that lean on hosted storage to keep costs down, such as the messaging service Twitter. But even the New York Times, which uses S3 to store articles from its historical archives, was spitting out internal server error messages.
Amazon has a pretty good track record in keeping its hosting services running. It's had some growing pains in the past, but this is the service's first major outage.
Service was restored a few hours later at 7:17 AM PST according to an Amazon technician.
"My business is effectively closed right now because Amazon did something wrong. I'll have to reconsider using the service now," wrote one customer on Amazon's web developer forum.
"This is the first outage I have experienced since I joined the service nearly a year ago. Yes it sucks, yes I hope they get it fixed very soon... but, the sky is not in fact falling at the moment," wrote another.
One likely outcome from Amazon's slip-up is some additional transparency in reporting future errors. S3 technicians were slow to confirm the problem and didn't provide any status updates to its customers. It took one hour to verify the outage from the first complaint on their developer board, and the S3 rep didn't post again until they resolved the issue.
"What really annoys me is the lack of an official explanation from Amazon (as usual)," posted one user.
"A health monitor would be useful - something to show what Amazon thinks the status of the services are and to post official information," wrote another.
Amazon said later that one of its three data centers for S3 was unreachable, beginning at 4:31 AM PST. It was back to "near normal" performance at 6:48 AM PST.
"Though we're proud of our uptime track record over the past two years with this service, any amount of downtime is unacceptable and we won't be satisfied until it's perfect," an Amazon rep posted.
An ultimately undeliverable "perfect" record may be the wrong message to be sending out. Maybe Amazon should start with convincing its users it'll do a better job than if they set up storage themselves. ®