xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-19 21:57:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

cleanup

Browse Source 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
This commit is contained in:
Robert Shaw 2025-07-20 17:04:35 +00:00
parent c08fb6d456
commit d9291f998e
1 changed files with 43 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

78
vllm/v1/metrics/loggers.py
View File

@ -425,29 +425,32 @@ class PrometheusStatLogger(StatLoggerBase):
self.histogram_decode_time_request = make_per_engine(
histogram_decode_time_request, engine_indexes, model_name)
# #
# # LoRA metrics
# #
#
# LoRA metrics
#
# # TODO: This metric might be incorrect in case of using multiple
# # api_server counts which uses prometheus mp.
# self.gauge_lora_info: Optional[prometheus_client.Gauge] = None
# if vllm_config.lora_config is not None:
# self.labelname_max_lora = "max_lora"
# self.labelname_waiting_lora_adapters = "waiting_lora_adapters"
# self.labelname_running_lora_adapters = "running_lora_adapters"
# self.max_lora = vllm_config.lora_config.max_loras
# self.gauge_lora_info = \
# self._gauge_cls(
# name="vllm:lora_requests_info",
# documentation="Running stats on lora requests.",
# multiprocess_mode="sum",
# labelnames=[
# self.labelname_max_lora,
# self.labelname_waiting_lora_adapters,
# self.labelname_running_lora_adapters,
# ],
# )
# TODO: This metric might be incorrect in case of using multiple
# api_server counts which uses prometheus mp.
self.gauge_lora_info: Optional[prometheus_client.Gauge] = None
if vllm_config.lora_config is not None:
if len(self.engine_indexes) > 1:
raise NotImplementedError(
"LoRA in DP mode is not supported yet.")
self.labelname_max_lora = "max_lora"
self.labelname_waiting_lora_adapters = "waiting_lora_adapters"
self.labelname_running_lora_adapters = "running_lora_adapters"
self.max_lora = vllm_config.lora_config.max_loras
self.gauge_lora_info = \
self._gauge_cls(
name="vllm:lora_requests_info",
documentation="Running stats on lora requests.",
multiprocess_mode="sum",
labelnames=[
self.labelname_max_lora,
self.labelname_waiting_lora_adapters,
self.labelname_running_lora_adapters,
],
)
def log_metrics_info(self, type: str, config_obj: SupportsMetricsInfo):
metrics_info = config_obj.metrics_info()
@ -471,7 +474,7 @@ class PrometheusStatLogger(StatLoggerBase):
for engine_index in self.engine_indexes:
metrics_info = config_obj.metrics_info()
metrics_info["engine"] = str(engine_index)
info_gauge.labels(*metrics_info.set(1))
info_gauge.labels(**metrics_info).set(1)
def record(self,
scheduler_stats: Optional[SchedulerStats],
@ -547,18 +550,21 @@ class PrometheusStatLogger(StatLoggerBase):
self.histogram_max_tokens_request[engine_idx].observe(
finished_request.max_tokens_param)
# if self.gauge_lora_info is not None:
# running_lora_adapters = \
# ",".join(iteration_stats.running_lora_adapters.keys())
# waiting_lora_adapters = \
# ",".join(iteration_stats.waiting_lora_adapters.keys())
# lora_info_labels = {
# self.labelname_running_lora_adapters: running_lora_adapters,
# self.labelname_waiting_lora_adapters: waiting_lora_adapters,
# self.labelname_max_lora: self.max_lora,
# }
# self.gauge_lora_info.labels(**lora_info_labels)\
# .set_to_current_time()
# TODO: investigate whether LoRA works with DP and ensure
# that the metrics we are logging are consistent with
if self.gauge_lora_info is not None:
running_lora_adapters = \
",".join(iteration_stats.running_lora_adapters.keys())
waiting_lora_adapters = \
",".join(iteration_stats.waiting_lora_adapters.keys())
lora_info_labels = {
self.labelname_running_lora_adapters: running_lora_adapters,
self.labelname_waiting_lora_adapters: waiting_lora_adapters,
self.labelname_max_lora: self.max_lora,
}
self.gauge_lora_info.labels(**lora_info_labels)\
.set_to_current_time()
def log_engine_initialized(self):
self.log_metrics_info("cache_config", self.vllm_config.cache_config)
@ -666,6 +672,8 @@ class StatLoggerManager:
logger.log()
def log_engine_initialized(self):
self.prometheus_logger.log_engine_initialized()
for per_engine_loggers in self.per_engine_logger_dict.values():
for logger in per_engine_loggers:
logger.log_engine_initialized()