- /api/dashboard now queries llm-api Prometheus metrics: req/s, tok/s, gen tok/s, uptime, sparklines - new LLM Inference card shows gen throughput + uptime badge + traffic sparkline - fmtUptime helper (human-readable duration)
- /api/dashboard now queries llm-api Prometheus metrics: req/s, tok/s, gen tok/s, uptime, sparklines - new LLM Inference card shows gen throughput + uptime badge + traffic sparkline - fmtUptime helper (human-readable duration)