How AWS infrastructure addresses Salesforce’s resiliency goals
summary
Salesforce describes concrete resiliency work on AWS: infrastructure as code with Terraform for immutable builds, containerized services on EKS with anti-affinity rules and health checks, and chaos engineering with steadybit to exercise concurrent failures. Operational practices include five READS metrics plus synthetic critical user journeys, an Average Page Time target <300 ms, AI that auto-detects 61% of incidents with mean detection <10 minutes and auto-remediates ~33% of customer incidents in <20 minutes, and immutable cross‑AZ storage plus DB driver failover.
AI-written summary · about
chapters (10)
- 0:00Introduction and overview
- 2:34Designing for fault domains
- 4:41Graceful degradation strategies
- 6:07Error budget and AI remediation
- 10:03Immutable infrastructure benefits
- 12:15Consistent metrics and CUJs
- 14:20EKS and Kubernetes resilience
- 16:25Chaos engineering approach
- 18:54Database resilience techniques
- 19:58Key takeaways and closing
related talks
32mA Peek Inside the Enterprise Cloud at Salesforce
23mJust enough DevOps for data scientists.
14mExpect the unexpected: Building resilience with AWS
57mResilient Cloud Architecture Design Patterns
31mDevOps in the Cloud: Case Studies of Amazon.com teams and their resilient architectures
35mBuilding Resilient Workloads on AWS: Tools and Techniques
26mA proactive approach to resilience in modern apps
26mService Ownership: DevOps for Salesforce