A production Claude system experiences sudden increase in response latency from 800ms average to 3,500ms average over a 2-hour period. The development team is receiving customer complaints about slow performance. The engineering lead needs to quickly diagnose whether the latency increase is caused by (1) increased traffic volume, (2) degraded model performance from Anthropic, (3) inefficient prompts consuming more tokens, or (4) database performance issues in context retrieval. Which operational enablement should have been implemented during system handoff to enable rapid diagnosis of this production performance issue?