Amazon scrambles to fix cloud networking issue affecting companies like Atlassian, Twilio

Key Points
  • Amazon Web Services acknowledged a networking issue at about 10:30 a.m. Eastern time on Friday.
  • Atlassian, Slack, Twilio and other web services reported problems that appeared to be related to the cloud issue.
Andy Jassy, CEO of Amazon Web Services and Infrastructure.
Source: Amazon Web Services

Amazon's cloud on Friday dealt with a technical issue in its important region of data centers in Virginia, and customers like Atlassian and Twilio found their services impacted.

Companies have been embracing public clouds from the likes of Amazon, Google and Microsoft in the past several years, and so when problems strike inside these companies' facilities, downtime for a variety of services can occasionally fan out across the internet.

Amazon first acknowledged the issue at around 10:30 a.m. Eastern time.

"We are investigating increased packet loss possibly impacting some AWS Direct Connect customers in the US-East-1 Region," the company said on the status page for its Amazon Web Services division.

Direct Connect, one of more than 100 services available from AWS, is a tool for sending data between AWS computing infrastructure and companies' existing data center equipment. The tool can "can reduce your network costs, increase bandwidth throughput, and provide a more consistent network experience than Internet-based connections," AWS says on its website.

After the initial notice, AWS indicated that the problems were going on across multiple data centers in US-East-1, which is large and has been available to customers for many years.

At 12:43 p.m. Eastern time, AWS said connections in CoreSite's VA1 and VA2 location in Reston, Va., and in Equinix's DC1-DC-6 and DC10-DC-12 location in Ashburn, Va., were "inactive."

"Inactive connections are not receiving routes advertised from Direct Connect routers," Amazon said. "We are now working to restore service on these Direct Connect connections. The AWS VPN service is operating normally and may be an alternative for some workloads."

One of the first web services to go down at the same time as the network disruption was Amazon's own voice assistant, Alexa, as TechCrunch reported. Amazon didn't immediately respond to a request for comment on whether the network issue was the cause of the Alexa problems.

Companies like Atlassian pointed to Amazon's Virginia infrastructure issue as the cause of connectivity issues with "most Atlassian products and services," including Bitbucket, Confluence and Jira.

GitHub, MongoDB, NewVoiceMedia, Slack and Zillow were among the companies that disclosed issues midday on Friday.

At 2:21 p.m. Eastern time AWS announced that Direct Connect was back to normal. "The issue has been resolved and the service is operating normally," AWS said.

On Monday, March 5, an AWS spokesman told CNBC in an email that the issues arose after a power loss at "one of AWS's redundant internet connection points in Virginia," which affected "a small number of AWS customers" using Direct Connect in the US-East-1 region.

"AWS resolved the issue and is working with its partner to prevent recurrence," the spokesman wrote.

A year ago AWS ran into challenges with its S3 storage service, affecting dozens of companies' services.

WATCH: Amazon cloud service to remain profitable

Amazon cloud business to remain profitable: Analyst
Amazon cloud business to remain profitable: Analyst