Amazon says Perplexity, Snapchat, and other services coming back soon, explains what went wrong

amazon

Amazon has issued a recovery update after a large-scale outage in its US-East-1 (N. Virginia) region disrupted several dependent services globally, including Perplexity AI, Snapchat, Reddit, Roblox, Prime Video, and others that rely on AWS infrastructure. The company says systems are showing signs of recovery and most requests are beginning to succeed, although residual failures and latency may continue until backlogs are cleared.

The incident affected multiple AWS products and global features that depend on the US-East-1 endpoint, including DynamoDB, IAM-related operations, support case creation, and global tables.

What Amazon says went wrong — root cause

According to Amazon’s official status timeline, engineers traced the issue to a DNS resolution failure affecting the DynamoDB API endpoint within the US-East-1 region. Since many AWS services and third-party applications depend on DynamoDB for core operations, the DNS failure triggered widespread request errors, timeouts, and service degradation across multiple platforms.

Timeline of AWS outage updates

Oct 20 — 12:11 AM PDT → 12:11 PM IST
Investigation begins into increased error rates and latency in US-East-1.

Oct 20 — 12:51 AM PDT → 12:51 PM IST
AWS confirms increased error rates and impact on support case creation.

Oct 20 — 1:26 AM PDT → 1:26 PM IST
Significant error rates confirmed for DynamoDB requests affecting other AWS services.

Oct 20 — 2:01 AM PDT → 2:01 PM IST
Root cause identified — DNS resolution issue on DynamoDB endpoint in US-East-1.

Oct 20 — 2:22 AM PDT → 2:22 PM IST
Initial mitigation applied; early recovery signs observed but failures continue.

Oct 20 — 2:27 AM PDT → 2:27 PM IST
Significant signs of recovery; majority of requests starting to succeed.

Oct 20 — 3:03 AM PDT → 3:03 PM IST
AWS confirms recovery across most affected services and global features tied to US-East-1.

Impact beyond AWS

Because US-East-1 is a critical region for hosting and routing workloads, the outage cascaded into consumer-facing services that run on AWS. Users of Perplexity AI, Snapchat, Reddit, Roblox, Prime Video and other platforms reported login failures, feed loading errors, media upload failures and server connection issues during the disruption window.

Recovery outlook

Amazon says the majority of affected services are recovering, but request retries may fail for some users until the backlog is cleared. No further degradation is currently reported, and Amazon will continue to issue updates until full restoration is confirmed.