llama.cpp server: a KV-cache utilization metric - VERIFIED ABSENT

VERIFIED NEGATIVE. llama.cpp server does not emit a KV-cache utilization metric. llamacpp_kv_cache_usage_perc is NOT a metric this engine registers - it is recorded here under the conventional name a reader would go looking for so the absence is queryable rather than implied by a gap in a list. Cache reuse IS observable as llamacpp:prompt_tokens_cached_total, so you can compute a hit rate - but occupancy is not exposed anywhere, so a multi-tenant deployment cannot tell how close to the context limit it is. What was searched: the to_metrics() tables contain no utilization or occupancy name; the only cache series is prompt_tokens_cached_total.. Do not read this record's metric_name as something the engine exports; it is the name that is missing.