SDK caching enables prompt access during API unavailability but fails on restart

Long-running services maintain access to prompts via SDK-level caching when Prompt Management API is unavailable. However, new deployments or application restarts cause downtime as they cannot fetch prompts from the unavailable API.

Ensure SDK cache TTL is configured appropriately for your deployment patterns. For critical services, consider pre-warming caches or maintaining warm standby instances. Implement graceful degradation strategies for new deployments during API outages. Monitor cache hit rates and API availability together.