In vLLM v0.23.0, vllm:prompt_tokens_cached increments from the iteration PromptTokenStats.cached_tokens field.

Metric

Applies when

Evidence

Does not establish

• a workload-independent healthy threshold
• semantics in any vLLM version other than v0.23.0

Guards

• vllm_version = v0.23.0
• metric_name = vllm:prompt_tokens_cached

Freshness

Pinned release-tag semantics; do not carry to another version without continuity evidence.