In vLLM v0.23.0, `queued_time` is computed as `req_stats.scheduled_ts` - `req_stats.queued_ts`; the source comment describes the queued interval as from first `QUEUED` event to first `SCHEDULED`.

MetricvLLM v0.23.0confidence: highverified 2 hr agoverified-code

V1 metricsIterationStats.update_from_finished_request finished-request accounting

vLLM v0.23.0 v1 metrics stats.py(vllm/v1/metrics/stats.py:439-440)

definesprimarysource_codecode_introspection

typical or healthy queue-time durations (workload-dependent)
that queued_ts reflects the first QUEUED event if a request emits multiple QUEUED events (IterationStats.update_from_events assigns queued_ts on each QUEUED event processed, so the source comment's "first QUEUED event" wording is not enforced by the visible assignment)
queue interval semantics in vLLM versions other than v0.23.0

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

In vLLM v0.23.0, queued_time is computed as req_stats.scheduled_ts - req_stats.queued_ts; the source comment describes the queued interval as from first QUEUED event to first SCHEDULED.