Update: At 10:27BST, a bit over two hours after the first report, AWS said it was "seeing significant signs of recovery".
"Most requests should now be succeeding. We continue to work through a backlog of queued requests," read its status update.
AWS said the root cause appeared to be DNS resolution of the DynamoDB API.
And it confirmed that some global services were still dependent on US-EAST-1.
See also our update here > AWS Outage: It’s always DNS, but… sometimes it’s overloaded network hardware
AWS confirmed "increased error rates and latencies" across a long list of services in the US-EAST-1 Region a little before 09:00 BST on Monday.
Some 30 minutes later, it said the problem appeared to be in requests to the DynamoDB endpoint.
To add insult to injury, the impact on Amazon's own services – apparently including store checkout – meant that some customers could not access the AWS support centre or its API to create cases. Again.
Various SaaS and other customers said they were impacted to at least some extent, with Xero and Asana both name-checking AWS. So did Coinbase and Perplexity, and Signal's troubles appeared related.
Other services and companies that either confirmed impact or saw degraded service at the exact same time included Snapchat, Ring, Jira, Slack, and some UK government departments.
Cloudflare reported some "delays in updating device posture information when using integrations with third-party posture providers", but did not name AWS. Docker likewise said it had "issue with one of our cloud service providers", without naming names.
AWS listed 17 services as affected, across identity and access management, Elastic Kubernetes, CloudFront and CloudWatch, and VPC Lattice.
Hello US-EAST-1, my old friend
The US-EAST-1 region has a long history of taking AWS support services down with it.
After two such cases in late 2021 – with some customers tremendously angry that their cloud provider had a single point of systemic risk – AWS shipped a new Service Health Dashboard on top of "an architecture that actively runs across multiple AWS regions to ensure we do not have delays in communicating with customers."
But, in June 2023, US-EAST-1 took down 100 services for four hours. Among the trouble: "customers may also have experienced issues when attempting to initiate a Call or Chat to AWS Support”.
In April, AWS announced a new Availability Zone in Maryland, with infrastructure located in US-EAST-1. It is due to go into general availability in 2026.
The link has been copied!
