Resilience Patterns Skill

External dependencies fail. Networks partition. Databases become slow. This skill covers the patterns that keep your service running when dependencies don't.

When to Activate

• Designing a service that calls external APIs, databases, or queues
• Adding retry logic to an HTTP client
• Implementing circuit breakers for external calls
• Reviewing architecture for single points of failure
• Preparing for a resilience audit or chaos experiment
• Deciding which HTTP error codes should be retried vs. treated as permanent failures
• Isolating slow dependencies using bulkhead thread-pool or connection-pool partitioning
• Configuring Kubernetes liveness, readiness, and startup probes for a new service

Circuit Breaker