/tag
Treating LLM calls as just another upstream dependency. How to use Spring AI to build a multi-provider gateway with retries, circuit breakers, prompt versioning, and observability — the same hygiene you'd put around any external API.