Schema.ai· vLLM· Schema 0.2.0
vLLM official site vLLM v0.23.0Log in

vLLM v0.16.0

knowledge available0 claims2 metric facts

Prefix-cache metrics

MetricStateMeaning
vllm:prefix_cache_hitspresentCounts prefix-cache activity in tokens.
vllm:prefix_cache_queriespresentCounts prefix-cache activity in tokens.