Performance & Scaling
Bug Report
LLM agent TTFT tripled: a timestamp in the system prompt broke prefix caching
After a harmless-looking prompt change, our agent's time to first token went from 0.5 s to 1.6 s and GPU prefill load tripled. vLLM's prefix cache hit rate had dropped from 87% to 2%. The cause was one line at the top of the system prompt, plus tool schemas serialised in random order.