vLLM prompt-token counter by source
vllm · counter
Fields
id
metric-exposures/w3m-vllm-vllm-prompt-tokens-by-sourceMetric name
vllm:prompt_tokens_by_source
Metric type
counter
Unit
tokens
Exposure
prometheus
Endpoint
:8000/metrics
Labels
model_nameenginesource
Enabled by default
yes
notes
vLLM main, PrometheusStatLogger in vllm/v1/metrics/loggers.py, read 2026-10-04. On by default (EngineArgs.disable_log_stats defaults False); served at :8000/metrics. Bucket family name in --custom-histogram-buckets: no bucket family. Splits prompt tokens by provenance (cache hit, external connector, multimodal cache).
Notes
vLLM main, PrometheusStatLogger in vllm/v1/metrics/loggers.py, read 2026-10-04. On by default (EngineArgs.disable_log_stats defaults False); served at :8000/metrics. Bucket family name in --custom-histogram-buckets: no bucket family. Splits prompt tokens by provenance (cache hit, external connector, multimodal cache).