resilience-patterns

Prevent cascading failures by implementing bulkhead isolation and timeout policies to contain failures and fail fast. This stops a single unresponsive dependency from exhausting application resources and impacting other services.

Handle API outages by implementing fallback strategies that provide alternative responses or actions. This enables degraded mode operation, allowing your application to return cached responses instead of crashing during external service failures.

Bulkhead isolation improves fault tolerance by limiting concurrent requests to prevent resource exhaustion, while hedged requests improve latency by issuing parallel calls and accepting the first available response.

Yes, you can implement concurrency limits and custom exception handling without external dependencies. The skill uses built-in asyncio semaphores and concurrent.futures to manage fault tolerance and robust error handling natively.

Use hedged requests for latency-sensitive operations where waiting on a single slow service degrades performance. This approach issues parallel requests and accepts the first response, optimizing overall application latency.

Yes, it works with AI providers like Anthropic and Ollama. When these external services experience downtime, the skill ensures your application remains stable by providing fallback responses and operating in a degraded mode.