Enhance Kubernetes high availability with Amazon Application Recovery Controller and Karpenter integration

This article explains how to integrate Amazon Application Recovery Controller (ARC) with Karpenter to enhance Kubernetes high availability during zone-level failures.

• ARC detects degraded Availability Zones and shifts traffic away automatically
• Karpenter lacks native integration with ARC, potentially provisioning capacity in impaired zones
• New open-source Kubernetes controller bridges ARC and Karpenter integration gap
• Controller reconfigures Karpenter node pools to exclude impaired zones during zonal shifts
• Uses EventBridge and SQS to monitor and process zonal autoshift events
• Automatically restores original node pool configuration when zone impairment resolves
• AWS FIS can trigger zonal autoshifts for chaos engineering and resilience testing

This integration ensures Karpenter's capacity provisioning aligns with ARC's traffic management during zone failures, improving application resilience without sacrificing efficiency.