Amazon Web Services said it is seeing “significant signs of recovery” after a major outage in its Northern Virginia (US-EAST-1) region early Monday disrupted multiple cloud services and downstream applications. The company traced the incident to DNS resolution problems affecting the DynamoDB API endpoint, according to the AWS Service Health Dashboard.
By around 3:35 a.m. PDT, AWS said the “underlying DNS issue” was fully mitigated and most operations were succeeding, while cautioning that some requests could be throttled and several services were working through event backlogs such as CloudTrail and Lambda, according to the AWS Service Health Dashboard.
New EC2 launches in US-EAST-1 continued to see elevated error rates into the early morning, and services that depend on EC2 — such as RDS, ECS and Glue — were affected. AWS advised customers to avoid pinning new instance launches to a single Availability Zone and to use multi-AZ Auto Scaling, according to the AWS Service Health Dashboard.
AWS also said Lambda’s event source mappings for SQS experienced polling delays, with teams working through SQS backlogs. By 5:10 a.m. PDT, the company said SQS queues processed via Lambda event source mappings had recovered and backlogs were being drained, according to the AWS Service Health Dashboard.
Earlier updates noted global services and features that rely on US-EAST-1 — including IAM updates and DynamoDB Global Tables — were impacted during the incident but had shown recovery as mitigations took effect, according to the AWS Service Health Dashboard.
The AWS Service Health Dashboard shows that AWS is working on additional fixes to restore full capacity for EC2 launches and continues to process backlogs across affected services.
A global media for the latest news, entertainment, music fashion, and more.
