Amazon Web Services Inc. is experiencing a technical issue with its Kinesis Data Streams service that has caused disruptions for Roku Inc., Adobe Systems Inc., Roomba maker iRobot Corp. and a number of other companies that rely on the cloud giant’s platform.

AWS said in an update posted at 12:15 p.m. PST that it’s making progress toward fixing the problem. Its engineers expect “full recovery to still take up to a few hours.”

Update: By early Thursday morning, Amazon said, the problem was solved.

“We have restored all traffic to Kinesis Data Streams via all endpoints, and have resolved the error rates invoking CloudWatch APIs,” the company said on an update on its AWS Service Health Dashboard. “All services are now operating normally. We have identified the root cause of the Kinesis Data Streams event, and have completed immediate actions to prevent recurrence.”

AWS customers use Kinesis Data Streams to transport information between their applications. On Wednesday morning at 6:36 a.m. PST, the cloud giant disclosed on its Service Health Dashboard web page that its engineers were “investigating increased error rates” for the service. The problems have been ongoing since then. 

The outage affects Kinesis Data Streams in one specific AWS data center cluster: the US-EAST-1 Region in Northern Virginia, which is made up of six availability zones. The cause is a malfunction in a subsystem responsible for handling incoming requests that is making it difficult to read and write information from the service. As a result, customers whose applications rely on Kinesis Data Streams to transport information have experienced problems.

The issue also affects certain other AWS products that rely on Kinesis Data Streams to power some of their features. EventBridge, another AWS service for transporting information between systems, was seeing “significant API faults and latencies” as of 10:44 a.m. PST. Many of the other affected products have been affected to a lesser extent: Certain EC2 instances, for example, displayed health metrics with delays at one point even though they weren’t encountering any problems.

“We continue to work towards recovery of the issue affecting the Kinesis Data Streams API in the US-EAST-1 Region,” AWS said in its most recent update on the Service Health Dashboard. “We also continue to see an improvement in error rates for Kinesis and several affected services.”

An Amazon AWS outage is currently impacting our iRobot Home App. Please know that our team is aware and monitoring the situation and hope to get the App back online soon. Thank you for your understanding and patience.

— iRobot (@iRobot) November 25, 2020

🚨We are aware of a service interruption impacting Roku accounts. We apologize for the inconvenience, find our status update at https://t.co/mxCNxoG36N.

— Roku Support (@RokuSupport) November 25, 2020

An Amazon AWS outage is currently impacting Adobe Spark so you may be having issues accessing/editing your projects. We are actively working with AWS and will report when the issue has subsided. https://t.co/uoHPf44HjL for current Spark status. We apologize for any inconvenience!

— Adobe Spark (@AdobeSpark) November 25, 2020

“The fact is, internet infrastructure can be unpredictable, especially when world events are driving unprecedented traffic and network congestion,” Kris Beevers, co-founder and chief executive of NS1 Inc., a Domain Name System and iInternet] traffic management technology company, told SiliconANGLE. “It is more important than ever to diversify infrastructure at every level. Using two or more cloud providers with distributed footprints ensures that if one provider experiences problems, automated failover systems can route traffic to a better-performing provider so that there is minimal impact to users.”

Photo: AWS