A number of popular Internet web sites and services hosted by Amazon Web Services
went offline or had various connectivity issues
beginning Sunday afternoon, with the problems quickly traced once again
to the company's northern Virginia data center, which has become notorious
for occasional uptime issues. The outage took down (at least for brief periods) Netflix
and others. All were back online within hours.Two sets of problems
- were uncovered in logs detailing the outage, which lasted up to five hours depending on the sites in question. Amazon reported at 4:29PM ET on Sunday that EBS volumes in its ECS database were having instance launch and network packet loss errors.
The latter issue was corrected about an hour later, but not all sites were fully restored until 9PM ET. In addition, the company also had issues with load balancers in the "US East 1" region. The load-balancing issue was first reported at 4:40PM and was identified and fixed about an hour later, but was only reported as fully resolved some 4.5 hours after the initial event.