In vLLM v0.23.0, vllm:request_inference_time_seconds is registered as a Prometheus histogram.
MetricvLLM v0.23.0confidence: highverified 1 hr agoverified-code
Metric
request_inference_time_secondsApplies when
PrometheusStatLogger metrics publisherV1 metrics
Evidence
vLLM v0.23.0 v1 metrics loggers.py(vllm/v1/metrics/loggers.py:890-895)
definesprimarysource_codecode_introspection
Does not establish
- the metric's unit of measure
- what the metric measures semantically
- a healthy or expected value range
Guards
vllm_version= v0.23.0metric_name=vllm:request_inference_time_seconds