All versions · Version range: v0.23.0

In vLLM v0.23.0, `vllm:inter_token_latency_seconds` is registered as a Prometheus histogram.

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

Metric

inter_token_latency_seconds

Applies when

PrometheusStatLogger metrics publisherV1 metrics

Evidence

vLLM v0.23.0 v1 metrics loggers.py(vllm/v1/metrics/loggers.py:787-812)

definesprimarysource_codecode_introspection

Open cited lines in immutable upstream source ↗

787:         histogram_inter_token_latency = self._histogram_cls(
788:             name="vllm:inter_token_latency_seconds",
789:             documentation="Histogram of inter-token latency in seconds.",
790:             buckets=[
791:                 0.01,
792:                 0.025,
793:                 0.05,
794:                 0.075,
795:                 0.1,
796:                 0.15,
797:                 0.2,
798:                 0.3,
799:                 0.4,
800:                 0.5,
801:                 0.75,
802:                 1.0,
803:                 2.5,
804:                 5.0,
805:                 7.5,
806:                 10.0,
807:                 20.0,
808:                 40.0,
809:                 80.0,
810:             ],
811:             labelnames=labelnames,
812:         )

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:inter_token_latency_seconds

Freshness

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