Explore resilient architecture patterns and chaos engineering methodologies in this comprehensive conference talk. Learn how modern software systems embrace failure instead of avoiding it, enhancing observability and leveraging techniques like graceful degradation, timeouts, and circuit breakers. Discover the benefits of intentionally breaking systems to build more resilient architectures. Dive into complex systems, partial failure modes, availability concepts, and AWS infrastructure. Examine workload isolation, cascading failures prevention, cell-based architectures, and shuffle sharding. Understand auto-scaling for self-healing, decoupling with async patterns, and traffic prioritization. Master essential practices such as setting timeouts, implementing backoff strategies, and designing idempotent operations. Explore health check types, service degradation techniques, and database optimization strategies. Gain insights into chaos engineering principles and their application in creating robust, fault-tolerant systems.
Overview
Syllabus
Intro
Complex systems
Partial failure mode
The Famous 95 of availability
Availability in parallel
AWS Region and availability zones
Multi-AZ architecture
Theoretical blast radius
Typical service application
Partial availability zone failure
System properties . Workload isolation
Cascading Failures
Cell-based architecture
Shuffle sharding
Auto-scaling for self-healing
Decoupling with async pattern
Degrade & prioritize traffic with queues
Set the timeouts!
Backing off between retries
Idempotent operation
Shallow health check
Deep health check
Service Degradation & Fallbacks
Pattern B: Circuit Breaker
Database Federation
Database Sharding
Read/Write separation
Pattem 9+
Chaos engineering
Taught by
NDC Conferences
