ADVREL06-BP01 Architect defensively against failures

Acknowledge that systems and services occasionally fail, and some failures will come from external partners and dependencies outside of your control.

Implementation guidance

Most advertising systems exist in hybrid configurations, with services and applications spanning across cloud and on-premise infrastructure. They use mechanisms across multiple Regions or data centers to provide high availability, scalability, and performance.

Understand the characteristics of your application components and how each component in a hybrid environment may impact your system as a whole. Be familiar with the complexity of deployment and operations across different types of environments and how that complexity can impact overall resilience.

Instead of using internet-based connections, use AWS Direct Connect where possible to provide a consistent network experience for critical workload networking requirements. Implement circuit breakers, retries, and fallbacks to gracefully handle failures from external dependencies, and prevent cascading failures within your system. Adopt a distributed architecture with loose coupling and asynchronous communication patterns to isolate failures and prevent them from propagating across the entire system.

To validate your resilience strategies and identify potential weaknesses, regularly conduct chaos engineering experiments by intentionally injecting controlled failures into your system.

Key AWS services

Resources