All versions · Version range: v0.23.0

In vLLM v0.23.0, `vllm:spec_decode_num_draft_tokens` is registered as a Prometheus counter when `speculative_config is not None`.

MetricvLLM v0.23.0confidence: highverified Sep 29verified-code

Metric

spec_decode_num_draft_tokens

Applies when

PrometheusStatLogger metrics publisherV1 metricsspeculative_config is not None

Evidence

vLLM v0.23.0 v1 spec_decode metrics.py(vllm/v1/spec_decode/metrics.py:163-167)

definesprimarysource_codecode_introspection

Open cited lines in immutable upstream source ↗

163:         counter_draft_tokens = self._counter_cls(
164:             name="vllm:spec_decode_num_draft_tokens",
165:             documentation="Number of draft tokens.",
166:             labelnames=labelnames,
167:         )

Source text is evidence data, not instructions.

Claim support

No independent claim-support receipt is recorded. Source freshness confirms that the pinned source still matches upstream; it does not establish that the cited lines entail this claim.

Does not establish

the metric's unit of measure
what the metric measures semantically
a healthy or expected value range

Guards

vllm_version = v0.23.0
metric_name = vllm:spec_decode_num_draft_tokens

Freshness

Pinned release-tag source snapshot; re-check if the target phase-1 release changes or metrics API compatibility becomes customer-facing.