vLLM metric identity

broadcovered at v0.23.0legacy evidence: v0.23.0evidence: source code

Coverage

Metric identity coverage anchored by the complete v0.23.0 Prometheus inventory, with multi-version history where an admitted scope has tracked a metric across releases.

Strong across the complete v0.23.0 fixed-name inventory; version-local coverage outside v0.23.0 is limited to metrics with admitted history records and must be shown as partial rather than inferred for the remaining inventory.

Generated slice, not hand-authored: every claim here is produced by scripts/generate_metric_identity_claims.py from registration call sites in the pinned v0.23.0 snapshots, and is reproducible by re-running it. Instrument-kind coverage is complete over the closed inventory (45 of 45); label and bucket coverage is partial by construction, and the shortfalls are declared below rather than inferred. Seven instrument-kind claims predate the generator, are hand-authored, and are checked against the extraction rather than overwritten by it.

Coverage at vLLM v0.23.0

Every declared topic entity has applicable claims, availability, or admitted history at this version.

94 claims2 history facts45/45 entities

History-backed metrics: prefix_cache_hits, prefix_cache_queries

Claims in this topic (94)

Claim	Verified
In vLLM v0.23.0, `vllm:cache_config_info` is registered as a Prometheus gauge.	11 hr ago
In vLLM v0.23.0, `vllm:corrupted_requests` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:corrupted_requests` is registered as a Prometheus counter when `VLLM_COMPUTE_NANS_IN_LOGITS=1`.	11 hr ago
In vLLM v0.23.0, `vllm:e2e_request_latency_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.3 to 7680.0.	11 hr ago
In vLLM v0.23.0, `vllm:e2e_request_latency_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:e2e_request_latency_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:engine_sleep_state` is registered with the label names `model_name`, engine, `sleep_state`.	11 hr ago
In vLLM v0.23.0, `vllm:engine_sleep_state` is registered as a Prometheus gauge.	11 hr ago
In vLLM v0.23.0, `vllm:estimated_flops_per_gpu_total` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:estimated_read_bytes_per_gpu_total` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:estimated_write_bytes_per_gpu_total` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:external_prefix_cache_hits` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:external_prefix_cache_hits` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:external_prefix_cache_queries` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:external_prefix_cache_queries` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:generation_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:generation_tokens` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:inter_token_latency_seconds` declares 19 literal histogram bucket boundaries in its registration, from 0.01 to 80.0.	11 hr ago
In vLLM v0.23.0, `vllm:inter_token_latency_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:inter_token_latency_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:iteration_tokens_total` declares 13 literal histogram bucket boundaries in its registration, from 1.0 to 16384.0.	11 hr ago
In vLLM v0.23.0, `vllm:iteration_tokens_total` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:iteration_tokens_total` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_idle_before_evict_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.001 to 1800.0.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_idle_before_evict_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_idle_before_evict_seconds` is registered as a Prometheus histogram when `observability_config.kv_cache_metrics=True`.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_lifetime_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.001 to 1800.0.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_lifetime_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_lifetime_seconds` is registered as a Prometheus histogram when `observability_config.kv_cache_metrics=True`.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_reuse_gap_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.001 to 1800.0.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_reuse_gap_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:kv_block_reuse_gap_seconds` is registered as a Prometheus histogram when `observability_config.kv_cache_metrics=True`.	11 hr ago
In vLLM v0.23.0, `vllm:kv_cache_usage_perc` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:kv_cache_usage_perc` is a Prometheus gauge for KV-cache usage where 1 means 100 percent usage.	11 hr ago
In vLLM v0.23.0, `vllm:lora_requests_info` is registered as a Prometheus gauge when `lora_config is not None`.	11 hr ago
In vLLM v0.23.0, `vllm:mm_cache_hits` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:mm_cache_hits` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:mm_cache_queries` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:mm_cache_queries` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:num_preemptions` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:num_preemptions` is a Prometheus counter for cumulative preemptions from the engine.	11 hr ago
In vLLM v0.23.0, `vllm:num_requests_running` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:num_requests_running` is a Prometheus gauge for the number of requests in model execution batches.	11 hr ago
In vLLM v0.23.0, `vllm:num_requests_waiting` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:num_requests_waiting` is a Prometheus gauge for the number of requests waiting to be processed.	11 hr ago
In vLLM v0.23.0, `vllm:num_requests_waiting_by_reason` is registered with the label names `model_name`, engine, reason.	11 hr ago
In vLLM v0.23.0, `vllm:num_requests_waiting_by_reason` is a Prometheus gauge with reason labels capacity and deferred, and its reasons sum to `vllm:num_requests_waiting`.	11 hr ago
In vLLM v0.23.0, `vllm:prefix_cache_hits` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:prefix_cache_hits` is a Prometheus counter measured in cached (hit) tokens — it counts tokens, not requests or cache blocks.	11 hr ago
In vLLM v0.23.0, `vllm:prefix_cache_queries` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:prefix_cache_queries` is a Prometheus counter measured in queried tokens — it counts tokens, not requests or cache blocks.	11 hr ago
In vLLM v0.23.0, `vllm:prompt_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:prompt_tokens` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:prompt_tokens_by_source` is registered with the label names `model_name`, engine, source.	11 hr ago
In vLLM v0.23.0, `vllm:prompt_tokens_by_source` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:prompt_tokens_cached` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:prompt_tokens_cached` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:request_decode_time_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.3 to 7680.0.	11 hr ago
In vLLM v0.23.0, `vllm:request_decode_time_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_decode_time_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_generation_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_generation_tokens` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_inference_time_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.3 to 7680.0.	11 hr ago
In vLLM v0.23.0, `vllm:request_inference_time_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_inference_time_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_max_num_generation_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_max_num_generation_tokens` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_params_max_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_params_max_tokens` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_params_n` declares 5 literal histogram bucket boundaries in its registration, from 1.0 to 20.0.	11 hr ago
In vLLM v0.23.0, `vllm:request_params_n` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_params_n` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_prefill_kv_computed_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_prefill_kv_computed_tokens` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_prefill_time_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.3 to 7680.0.	11 hr ago
In vLLM v0.23.0, `vllm:request_prefill_time_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_prefill_time_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_prompt_tokens` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_prompt_tokens` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_queue_time_seconds` declares 21 literal histogram bucket boundaries in its registration, from 0.3 to 7680.0.	11 hr ago
In vLLM v0.23.0, `vllm:request_queue_time_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_queue_time_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:request_success` is registered with the label names `model_name`, engine, `finished_reason`.	11 hr ago
In vLLM v0.23.0, `vllm:request_success` is registered as a Prometheus counter.	11 hr ago
In vLLM v0.23.0, `vllm:request_time_per_output_token_seconds` declares 19 literal histogram bucket boundaries in its registration, from 0.01 to 80.0.	11 hr ago
In vLLM v0.23.0, `vllm:request_time_per_output_token_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:request_time_per_output_token_seconds` is registered as a Prometheus histogram.	11 hr ago
In vLLM v0.23.0, `vllm:spec_decode_num_accepted_tokens` is registered as a Prometheus counter when `speculative_config is not None`.	11 hr ago
In vLLM v0.23.0, `vllm:spec_decode_num_accepted_tokens_per_pos` is registered as a Prometheus counter when `speculative_config is not None`.	11 hr ago
In vLLM v0.23.0, `vllm:spec_decode_num_draft_tokens` is registered as a Prometheus counter when `speculative_config is not None`.	11 hr ago
In vLLM v0.23.0, `vllm:spec_decode_num_drafts` is registered as a Prometheus counter when `speculative_config is not None`.	11 hr ago
In vLLM v0.23.0, `vllm:time_to_first_token_seconds` declares 22 literal histogram bucket boundaries in its registration, from 0.001 to 2560.0.	11 hr ago
In vLLM v0.23.0, `vllm:time_to_first_token_seconds` is registered with the label names `model_name`, engine.	11 hr ago
In vLLM v0.23.0, `vllm:time_to_first_token_seconds` is registered as a Prometheus histogram.	11 hr ago

Metrics in this topic (45)

cache_config_info — 1 claim
corrupted_requests — 2 claims
e2e_request_latency_seconds — 3 claims
engine_sleep_state — 2 claims
estimated_flops_per_gpu_total — 1 claim
estimated_read_bytes_per_gpu_total — 1 claim
estimated_write_bytes_per_gpu_total — 1 claim
external_prefix_cache_hits — 2 claims
external_prefix_cache_queries — 2 claims
generation_tokens — 2 claims
inter_token_latency_seconds — 3 claims
iteration_tokens_total — 3 claims
kv_block_idle_before_evict_seconds — 3 claims
kv_block_lifetime_seconds — 3 claims
kv_block_reuse_gap_seconds — 3 claims
kv_cache_usage_perc — 2 claims
lora_requests_info — 1 claim
mm_cache_hits — 2 claims
mm_cache_queries — 2 claims
num_preemptions — 2 claims
num_requests_running — 2 claims
num_requests_waiting — 2 claims
num_requests_waiting_by_reason — 2 claims
prefix_cache_hits — 2 claims
prefix_cache_queries — 2 claims
prompt_tokens — 2 claims
prompt_tokens_by_source — 2 claims
prompt_tokens_cached — 2 claims
request_decode_time_seconds — 3 claims
request_generation_tokens — 2 claims
request_inference_time_seconds — 3 claims
request_max_num_generation_tokens — 2 claims
request_params_max_tokens — 2 claims
request_params_n — 3 claims
request_prefill_kv_computed_tokens — 2 claims
request_prefill_time_seconds — 3 claims
request_prompt_tokens — 2 claims
request_queue_time_seconds — 3 claims
request_success — 2 claims
request_time_per_output_token_seconds — 3 claims
spec_decode_num_accepted_tokens — 1 claim
spec_decode_num_accepted_tokens_per_pos — 1 claim
spec_decode_num_draft_tokens — 1 claim
spec_decode_num_drafts — 1 claim
time_to_first_token_seconds — 3 claims

Surfaces

Prometheus /metrics endpoint — 45 inventoried members

Known gaps

The corpus records no unit of measure for any v0.23.0 metric. vLLM's registration code has no structured unit field — `unit=` appears zero times in the pinned snapshots — so the only in-code signal is prose inside the `documentation=` argument, which states a unit for a minority of metrics and would require inference for the rest.

Schema.ai can state a metric's Prometheus instrument kind but cannot answer "what unit is this measured in" from structured corpus data. Where a registration docstring states a unit, it may appear in a claim statement as source-faithful prose, but it is not a queryable field.

Check: Admit a has_unit predicate only together with a rule for what grounds a unit — registration prose, the metrics design doc, or neither. Inferring from name suffixes such as `_seconds` is explicitly not grounded enough.

Five histogram metrics declare bucket boundaries computed at runtime via `build_1_2_5_buckets(max_model_len)` rather than as literal values; the corpus records no boundaries for them. Thirteen of the eighteen bucket declarations are literal and are recorded.

Schema.ai can list bucket boundaries for the thirteen histograms that declare them literally and must decline for the five computed ones. Asserting fixed boundaries for a set derived from a deployment's max_model_len would be false.

Check: Admit computed_from, or an equivalent way to state a derivation and its inputs, before recording any bucket fact for these five metrics.

Nine metrics build their label set at runtime rather than declaring it statically — from a constructor parameter, from `metrics_info.keys()`, or from instance attributes — so the corpus records no label names for them. Thirty-six of the forty-five have statically resolvable label sets.

Schema.ai can list label names for thirty-six metrics and must decline for the other nine, rather than guessing a label set from the common `model_name`/`engine` pattern.

Check: Resolving these requires reading construction sites outside the registration call. Decide whether cross-call-site resolution is in scope before recording labels for them.