NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·2h ago
Cloud & Infrastructure
The New Stack·25d ago·5 min read
What running Kubernetes across millions of clusters taught AWS about zone failures
TL;DR
AWS learned that handling subtle zone failures, where a zone is slow but not completely down, can lead to regional outages. To address this, they built zonal resiliency into Amazon EKS, focusing on static stability during impairments.
✦ Why It Matters
Engineers should implement static stability principles in their own Kubernetes deployments to enhance resilience against zone failures.
Key Takeaways
- Enable zonal shift to protect workloads from zone failures.
- Spread workloads across multiple Availability Zones for resilience.
- Test your application’s ability to handle zonal failures regularly.
- Pre-scale resources to ensure healthy zones can absorb traffic.
How It Works
The automated weight-shift mechanism in Amazon EKS quickly reroutes control plane activity from impaired zones by failing health checks for those zones, allowing traffic to be directed to healthy zones. This is achieved through DNS updates, scaling group adjustments, and leadership transfers among Kubernetes controllers, ensuring that the system remains operational and responsive.