Build Observable AI Model Fallbacks with Telnyx AI Gateway

Configure policy-driven LLM retries and fallback models, force a failover, and verify which model served the request through response headers and the spend ledger.