Patterns for resilient, highly available cloud architectures
summary
Covers practical cloud resilience patterns: deploy across multiple AWS regions and availability zones, use auto-scaling, replicas (Aurora) and asynchronous queues with worker push for background jobs, and make operations idempotent. Recommends exponential backoff with jitter for retries, circuit breakers (e.g. Hystrix), shuffle sharding for client isolation, load shedding, and running chaos experiments with AWS Fault Injection Simulator to validate recovery.
AI-written summary · about
chapters (18)
- 0:00Start
- 0:48There's always something that doesn't work
- 1:35Most systems are more monolithic
- 2:56Resilience example
- 3:21Design patterns catalog
- 5:21Geography
- 7:20Availability zones
- 10:27Autoscaling
- 11:17Asynchronous
- 14:28Databases
- 16:49Timeout management
- 20:05How to avoid timeouts?
- 25:21How to test an application?
- 27:20Message queues
- 31:36Sharding
- 37:23Chaos Engineering
- 42:24Conclusion
- 44:26Q&A
related talks
57mResilient Cloud Architecture Design Patterns
01h 03mPatterns for Resilient Architecture
35mDesign patterns for application resiliency
53mResiliency and Availability Design Patterns for the Cloud
51mPatterns for building resilient software systems
35mBuilding Resilient Workloads on AWS: Tools and Techniques
45mA resilient application in a partially degraded world
37mResilience Design Patterns