What it means
Architectures use redundancy, diversity, graceful degradation, isolation, capacity, recovery automation, and tested operating procedures to manage failure.
Why it matters
Availability depends on complete service chains, not simply having a duplicate server or data copy.
Practical focus
- Identify critical dependencies and single points of failure
- Design for partial failure and degraded operation
- Separate redundant components across failure domains
- Test failover and recovery under realistic conditions
Common mistakes
- Calling duplicate components resilient without testing
- Replicating corruption or misconfiguration
- Ignoring people, identity, DNS, and supplier dependencies
Certification relevance
This subject appears in or supports the following certification bodies of knowledge: