Resilience Patterns
Implements production-grade fault tolerance and resilience patterns for distributed systems and LLM-based workflows.
The Resilience Patterns skill provides Claude Code with the architectural expertise needed to build robust, fault-tolerant distributed systems. It facilitates the implementation of industry-standard patterns such as circuit breakers to prevent cascading failures, bulkhead isolation to partition resources, and intelligent retry strategies with exponential backoff. Designed for modern AI development, it includes specialized logic for LLM API resilience, including model fallback chains and token budget management, ensuring that your AI agents and backend services remain reliable under load or during third-party outages.
主要功能
01Token budget guarding to manage context limits and optimize costs
02Bulkhead isolation patterns to prevent resource exhaustion across system tiers
03Advanced retry logic featuring exponential backoff and randomized jitter
04Circuit Breaker implementation with configurable failure thresholds and recovery timeouts
05LLM-specific fallback chains for multi-model reliability and graceful degradation
0669 GitHub stars
使用场景
01Protecting distributed microservices from cascading failures using circuit breakers
02Optimizing LLM integrations with automated fallbacks and smart error classification
03Developing high-availability multi-agent systems that handle API rate limits and model outages