In this episode of AWS Partners Unplugged, Tarek Makota, Senior Principal Solution Architect at AWS, discusses building resilient SaaS applications for ISV partners. We explore AWS fault isolation boundaries through regions and availability zones, explaining how proper application design can contain failures and prevent them from spreading across an entire customer base. We discuss the misconception that multi-region deployments automatically provide higher resilience, emphasizing that single-region architectures can achieve high availability when properly designed to honor fault boundaries between control planes and data planes. We cover key strategies including load shedding, careful retry implementation to avoid amplifying problems, identifying critical user journeys, and leveraging AWS services like Application Recovery Controller and its Zonal Auto Shift feature, along with Fault Injection Simulator for testing and recovery.

Links:

• https://aws.amazon.com/resilience-hub/
• https://docs.aws.amazon.com/prescriptive-guidance/latest/resilience-analysis-framework/introduction.html
• https://docs.aws.amazon.com/prescriptive-guidance/latest/resilience-lifecycle-framework/introduction.html

Contact us:

• aws-partner-unplugged-podcast-hosts@amazon.com
• Tarik: https://www.linkedin.com/in/tarik-makota/