LMDeploy end-to-end request latency histogram

InternLM/lmdeploy main, PrometheusStatLogger in lmdeploy/metrics/loggers.py, read 2026-10-04. On by default and requires NO flag: the endpoint is simply available at the API server port (23333 by default). Label keys are model_name and engine on every metric, where engine carries the dp_rank. Bucket construction reuses vLLM's build_buckets / build_1_2_5_buckets idiom. Shared request_latency ladder. The whole per-request phase set is observed in record_finish(); the prefill/decode split is what makes this set phase-scoped even though TTFT/TPOT are observed per iteration.