Building Resilient Systems
summary
A recent global outage eight months ago cascades into months of recovery and an estimated $5.5B estimated loss, with incidents like 14 production lines halted over a disk-space bug and a major social network losing internet for six hours; a hyperscaler also wiped customer data via an automation rule. The practical response is threefold: build redundant architecture with auto failovers and circuit breakers and AIops, formalize recovery playbooks and DR tiers, and validate via chaos engineering live-fail drills and layered maturity metrics.
AI-written summary · about
chapters (9)
- 0:00Introduction to outage scenario
- 3:13Why resiliency is critical now
- 7:45Weak links in modern systems
- 11:56Three pillars of resiliency
- 15:29Challenges in resiliency journey
- 20:07Layered approach to resiliency
- 22:53Measuring resiliency progress
- 26:05Testing resiliency with drills
- 28:37Conclusion and call to action
related talks
27mBuilding Ultra-Resilient Systems in Go
10mBuilding an 'Oops-Less Operation' at Scale
01h 03mPatterns for Resilient Architecture
53mResiliency and Availability Design Patterns for the Cloud
30mWhat Breaks Our Systems: A Taxonomy of Black Swans
29mCulturing Resiliency with Data: A Taxonomy of Outages
18mEngineering Resilience at Scale; Lessons from a Distributed Systems Transformation
55mGoing Resilient