SGLang pool capacity gauge
sglang · gauge
Fields
id
metric-exposures/w3m-sglang-sglang-max-total-num-tokensMetric name
sglang:max_total_num_tokens
Metric type
gauge
Unit
tokens
Exposure
prometheus
Endpoint
:30000/metrics
Labels
model_nameengine_typetp_rankpp_rankmoe_ep_rank
Enabled by default
no
notes
SGLang main, python/sglang/srt/observability/metrics_collector.py, read 2026-10-04. NOT ON BY DEFAULT: nothing is exported unless --enable-metrics (ServerArgs default False), which is also the only thing that mounts /metrics on the server port. Label keys are model_name, engine_type, tp_rank, pp_rank, moe_ep_rank (plus dp_rank under data parallelism, priority under priority scheduling, and anything in extra_metric_labels). Bucket ladders are overridable per histogram via get_histogram_conf_from_env. The denominator for any token-usage ratio.
Notes
SGLang main, python/sglang/srt/observability/metrics_collector.py, read 2026-10-04. NOT ON BY DEFAULT: nothing is exported unless --enable-metrics (ServerArgs default False), which is also the only thing that mounts /metrics on the server port. Label keys are model_name, engine_type, tp_rank, pp_rank, moe_ep_rank (plus dp_rank under data parallelism, priority under priority scheduling, and anything in extra_metric_labels). Bucket ladders are overridable per histogram via get_histogram_conf_from_env. The denominator for any token-usage ratio.