AWS

🔴 EC2 instances in use1-az4 decided to self-immolate after hitting thermal limits. AWS is playing server firefighter while customers watch their infrastructure turn into a convection oven. EBS volumes are also having a heat-induced existential crisis.

⏰ May 08, 2026 Fixed (for now) spotted via status_rss

// The official story

Service impact: Increased Error Rate and Latency

// Status page says

We continue working to resolve the impaired EC2 instances and degraded EBS volumes in a single Availability Zone (use1-az4) in the US-EAST-1 Region caused by a thermal event. During such an event, servers automatically shut down when the temperatures exceeded the operating thresholds in order to protect the hardware. We are actively working to bring additional cooling system capacity online, which will enable us to recover the remaining affected hardware in the impacted zone. Some customers will

🔗 Official status page (for the PR-approved version)

https://status.aws.amazon.com/

Aws's Greatest Hits

🟢 AWS Aug 18, 2026

⚪ AWS Direct Connect got surprise baptism at Equinix Frankfurt when water decided cooling systems weren't its only calling. Network gear drowned faster than a MacBook in a bathtub, took out packet delivery service for several hours. Plot twist: the real disaster was the infrastructure, not the cloud. (Auto-resolved: no updates in 24h)

🟢 AWS Aug 16, 2026

🔴 AWS's Frankfurt facility discovered that packets, like tourists, sometimes prefer getting lost. Co-location partner and AWS are having a spirited disagreement about how 'near term' recovery actually works. Spoiler: it's not. (Auto-resolved: no updates in 24h)

🟢 AWS Jul 31, 2026

⚪ AWS's route-assignment system forgot how to assign routes, turning AP-SOUTH-1 into a packet massacre zone. Someone's config change turned into someone else's 2 AM incident call. Fixed now, but the damage to your SLA is already in the group chat. (Auto-resolved: no updates in 24h)

🟢 AWS Jul 24, 2026

⚪ 🔌 AWS US-WEST-2 ghosted everyone for 20 minutes between 3:55-4:15 AM PDT, tanking connectivity across multiple services and the Management Console. Engineers showed up fashionably late at 4:01 AM, found the problem, and fixed it before anyone could file a ticket. The region's internal plumbing stayed dry though, so there's that. (Auto-resolved: no updates in 24h)