[ Bugfix ] Fix Prometheus Metrics With zeromq Frontend (#7279)

Co-authored-by: Nick Hill <nickhill@us.ibm.com>

tests/entrypoints/openai/test_basic.py

@@ -50,12 +50,3 @@ async def test_check_health(client: openai.AsyncOpenAI):
     response = requests.get(base_url + "/health")
 
     assert response.status_code == HTTPStatus.OK
-
-
-@pytest.mark.asyncio
-async def test_log_metrics(client: openai.AsyncOpenAI):
-    base_url = str(client.base_url)[:-3].strip("/")
-
-    response = requests.get(base_url + "/metrics")
-
-    assert response.status_code == HTTPStatus.OK

tests/entrypoints/openai/test_metrics.py

@@ -0,0 +1,179 @@
+from http import HTTPStatus
+
+import openai
+import pytest
+import requests
+from prometheus_client.parser import text_string_to_metric_families
+from transformers import AutoTokenizer
+
+from ...utils import RemoteOpenAIServer
+
+MODEL_NAME = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
+
+
+@pytest.fixture(scope="module")
+def default_server_args():
+    return [
+        # use half precision for speed and memory savings in CI environment
+        "--dtype",
+        "bfloat16",
+        "--max-model-len",
+        "1024",
+        "--enforce-eager",
+        "--max-num-seqs",
+        "128",
+    ]
+
+
+@pytest.fixture(scope="module",
+                params=[
+                    "",
+                    "--enable-chunked-prefill",
+                    "--disable-frontend-multiprocessing",
+                ])
+def client(default_server_args, request):
+    if request.param:
+        default_server_args.append(request.param)
+    with RemoteOpenAIServer(MODEL_NAME, default_server_args) as remote_server:
+        yield remote_server.get_async_client()
+
+
+_PROMPT = "Hello my name is Robert and I love magic"
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+_TOKENIZED_PROMPT = tokenizer(_PROMPT)["input_ids"]
+
+_NUM_REQUESTS = 10
+_NUM_PROMPT_TOKENS_PER_REQUEST = len(_TOKENIZED_PROMPT)
+_NUM_GENERATION_TOKENS_PER_REQUEST = 10
+
+# {metric_family: [(suffix, expected_value)]}
+EXPECTED_VALUES = {
+    "vllm:time_to_first_token_seconds": [("_count", _NUM_REQUESTS)],
+    "vllm:time_per_output_token_seconds":
+    [("_count", _NUM_REQUESTS * (_NUM_GENERATION_TOKENS_PER_REQUEST - 1))],
+    "vllm:e2e_request_latency_seconds": [("_count", _NUM_REQUESTS)],
+    "vllm:request_prompt_tokens":
+    [("_sum", _NUM_REQUESTS * _NUM_PROMPT_TOKENS_PER_REQUEST),
+     ("_count", _NUM_REQUESTS)],
+    "vllm:request_generation_tokens":
+    [("_sum", _NUM_REQUESTS * _NUM_GENERATION_TOKENS_PER_REQUEST),
+     ("_count", _NUM_REQUESTS)],
+    "vllm:request_params_n": [("_count", _NUM_REQUESTS)],
+    "vllm:request_params_best_of": [("_count", _NUM_REQUESTS)],
+    "vllm:prompt_tokens": [("_total",
+                            _NUM_REQUESTS * _NUM_PROMPT_TOKENS_PER_REQUEST)],
+    "vllm:generation_tokens":
+    [("_total", _NUM_REQUESTS * _NUM_PROMPT_TOKENS_PER_REQUEST)],
+    "vllm:request_success": [("_total", _NUM_REQUESTS)],
+}
+
+
+@pytest.mark.asyncio
+async def test_metrics_counts(client: openai.AsyncOpenAI):
+    base_url = str(client.base_url)[:-3].strip("/")
+
+    for _ in range(_NUM_REQUESTS):
+        # sending a request triggers the metrics to be logged.
+        await client.completions.create(
+            model=MODEL_NAME,
+            prompt=_TOKENIZED_PROMPT,
+            max_tokens=_NUM_GENERATION_TOKENS_PER_REQUEST)
+
+    response = requests.get(base_url + "/metrics")
+    print(response.text)
+    assert response.status_code == HTTPStatus.OK
+
+    # Loop over all expected metric_families
+    for metric_family, suffix_values_list in EXPECTED_VALUES.items():
+        found_metric = False
+
+        # Check to see if the metric_family is found in the prom endpoint.
+        for family in text_string_to_metric_families(response.text):
+            if family.name == metric_family:
+                found_metric = True
+
+                # Check that each suffix is found in the prom endpoint.
+                for suffix, expected_value in suffix_values_list:
+                    metric_name_w_suffix = f"{metric_family}{suffix}"
+                    found_suffix = False
+
+                    for sample in family.samples:
+                        if sample.name == metric_name_w_suffix:
+                            found_suffix = True
+
+                            # For each suffix, value sure the value matches
+                            # what we expect.
+                            assert sample.value == expected_value, (
+                                f"{metric_name_w_suffix} expected value of "
+                                f"{expected_value} did not match found value "
+                                f"{sample.value}")
+                            break
+                    assert found_suffix, (
+                        f"Did not find {metric_name_w_suffix} in prom endpoint"
+                    )
+                break
+
+        assert found_metric, (f"Did not find {metric_family} in prom endpoint")
+
+
+EXPECTED_METRICS = [
+    "vllm:num_requests_running",
+    "vllm:num_requests_swapped",
+    "vllm:num_requests_waiting",
+    "vllm:gpu_cache_usage_perc",
+    "vllm:cpu_cache_usage_perc",
+    "vllm:time_to_first_token_seconds_sum",
+    "vllm:time_to_first_token_seconds_bucket",
+    "vllm:time_to_first_token_seconds_count",
+    "vllm:time_per_output_token_seconds_sum",
+    "vllm:time_per_output_token_seconds_bucket",
+    "vllm:time_per_output_token_seconds_count",
+    "vllm:e2e_request_latency_seconds_sum",
+    "vllm:e2e_request_latency_seconds_bucket",
+    "vllm:e2e_request_latency_seconds_count",
+    "vllm:request_prompt_tokens_sum",
+    "vllm:request_prompt_tokens_bucket",
+    "vllm:request_prompt_tokens_count",
+    "vllm:request_generation_tokens_sum",
+    "vllm:request_generation_tokens_bucket",
+    "vllm:request_generation_tokens_count",
+    "vllm:request_params_n_sum",
+    "vllm:request_params_n_bucket",
+    "vllm:request_params_n_count",
+    "vllm:request_params_best_of_sum",
+    "vllm:request_params_best_of_bucket",
+    "vllm:request_params_best_of_count",
+    "vllm:num_preemptions_total",
+    "vllm:prompt_tokens_total",
+    "vllm:generation_tokens_total",
+    "vllm:request_success_total",
+    "vllm:cache_config_info",
+    # labels in cache_config_info
+    "block_size",
+    "cache_dtype",
+    "cpu_offload_gb",
+    "enable_prefix_caching",
+    "gpu_memory_utilization",
+    "num_cpu_blocks",
+    "num_gpu_blocks",
+    "num_gpu_blocks_override",
+    "sliding_window",
+    "swap_space_bytes",
+]
+
+
+@pytest.mark.asyncio
+async def test_metrics_exist(client: openai.AsyncOpenAI):
+    base_url = str(client.base_url)[:-3].strip("/")
+
+    # sending a request triggers the metrics to be logged.
+    await client.completions.create(model=MODEL_NAME,
+                                    prompt="Hello, my name is",
+                                    max_tokens=5,
+                                    temperature=0.0)
+
+    response = requests.get(base_url + "/metrics")
+    assert response.status_code == HTTPStatus.OK
+
+    for metric in EXPECTED_METRICS:
+        assert metric in response.text

vllm/engine/async_llm_engine.py

@@ -15,7 +15,7 @@ from vllm.engine.arg_utils import AsyncEngineArgs
 from vllm.engine.async_timeout import asyncio_timeout
 from vllm.engine.llm_engine import (DecoderPromptComponents, LLMEngine,
                                     PromptComponents)
-from vllm.engine.metrics import StatLoggerBase
+from vllm.engine.metrics_types import StatLoggerBase
 from vllm.executor.executor_base import ExecutorAsyncBase
 from vllm.executor.ray_utils import initialize_ray_cluster, ray
 from vllm.inputs import (EncoderDecoderLLMInputs, LLMInputs, PromptInputs,

vllm/engine/llm_engine.py

@@ -16,8 +16,7 @@ from vllm.config import (CacheConfig, DecodingConfig, DeviceConfig,
 from vllm.core.scheduler import (ScheduledSequenceGroup, Scheduler,
                                  SchedulerOutputs)
 from vllm.engine.arg_utils import EngineArgs
-from vllm.engine.metrics import (LoggingStatLogger, PrometheusStatLogger,
+from vllm.engine.metrics_types import StatLoggerBase, Stats
-                                 StatLoggerBase, Stats)
 from vllm.engine.output_processor.interfaces import (
     SequenceGroupOutputProcessor)
 from vllm.engine.output_processor.stop_checker import StopChecker
@@ -339,6 +338,13 @@ class LLMEngine:
             if stat_loggers is not None:
                 self.stat_loggers = stat_loggers
             else:
+                # Lazy import for prometheus multiprocessing.
+                # We need to set PROMETHEUS_MULTIPROC_DIR environment variable
+                # before prometheus_client is imported.
+                # See https://prometheus.github.io/client_python/multiprocess/
+                from vllm.engine.metrics import (LoggingStatLogger,
+                                                 PrometheusStatLogger)
+
                 self.stat_loggers = {
                     "logging":
                     LoggingStatLogger(

vllm/engine/metrics.py

@@ -1,13 +1,12 @@
-import time
-from abc import ABC, abstractmethod
-from dataclasses import dataclass
 from typing import TYPE_CHECKING
 from typing import Counter as CollectionsCounter
-from typing import Dict, List, Optional, Protocol, Union
+from typing import Dict, List, Optional, Union
 
 import numpy as np
 import prometheus_client
 
+from vllm.engine.metrics_types import (StatLoggerBase, Stats,
+                                       SupportsMetricsInfo)
 from vllm.executor.ray_utils import ray
 from vllm.logger import init_logger
 
@@ -29,41 +28,49 @@ prometheus_client.disable_created_metrics()
 
 # begin-metrics-definitions
 class Metrics:
+    """
+    vLLM uses a multiprocessing-based frontend for the OpenAI server.
+    This means that we need to run prometheus_client in multiprocessing mode
+    See https://prometheus.github.io/client_python/multiprocess/ for more
+    details on limitations.
+    """
     labelname_finish_reason = "finished_reason"
     _gauge_cls = prometheus_client.Gauge
     _counter_cls = prometheus_client.Counter
     _histogram_cls = prometheus_client.Histogram
 
     def __init__(self, labelnames: List[str], max_model_len: int):
-        # Unregister any existing vLLM collectors
+        # Unregister any existing vLLM collectors (for CI/CD)
         self._unregister_vllm_metrics()
 
-        # Config Information
-        self._create_info_cache_config()
-
         # System stats
         #   Scheduler State
         self.gauge_scheduler_running = self._gauge_cls(
             name="vllm:num_requests_running",
             documentation="Number of requests currently running on GPU.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
         self.gauge_scheduler_waiting = self._gauge_cls(
             name="vllm:num_requests_waiting",
             documentation="Number of requests waiting to be processed.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
         self.gauge_scheduler_swapped = self._gauge_cls(
             name="vllm:num_requests_swapped",
             documentation="Number of requests swapped to CPU.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
         #   KV Cache Usage in %
         self.gauge_gpu_cache_usage = self._gauge_cls(
             name="vllm:gpu_cache_usage_perc",
             documentation="GPU KV-cache usage. 1 means 100 percent usage.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
         self.gauge_cpu_cache_usage = self._gauge_cls(
             name="vllm:cpu_cache_usage_perc",
             documentation="CPU KV-cache usage. 1 means 100 percent usage.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
 
         # Iteration stats
         self.counter_num_preemption = self._counter_cls(
@@ -137,11 +144,13 @@ class Metrics:
         self.gauge_spec_decode_draft_acceptance_rate = self._gauge_cls(
             name="vllm:spec_decode_draft_acceptance_rate",
             documentation="Speulative token acceptance rate.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
         self.gauge_spec_decode_efficiency = self._gauge_cls(
             name="vllm:spec_decode_efficiency",
             documentation="Speculative decoding system efficiency.",
-            labelnames=labelnames)
+            labelnames=labelnames,
+            multiprocess_mode="sum")
         self.counter_spec_decode_num_accepted_tokens = (self._counter_cls(
             name="vllm:spec_decode_num_accepted_tokens_total",
             documentation="Number of accepted tokens.",
@@ -160,19 +169,18 @@ class Metrics:
             name="vllm:avg_prompt_throughput_toks_per_s",
             documentation="Average prefill throughput in tokens/s.",
             labelnames=labelnames,
+            multiprocess_mode="sum",
         )
         # Deprecated in favor of vllm:generation_tokens_total
         self.gauge_avg_generation_throughput = self._gauge_cls(
             name="vllm:avg_generation_throughput_toks_per_s",
             documentation="Average generation throughput in tokens/s.",
             labelnames=labelnames,
+            multiprocess_mode="sum",
         )
 
-    def _create_info_cache_config(self) -> None:
+
-        # Config Information
+# end-metrics-definitions
-        self.info_cache_config = prometheus_client.Info(
-            name='vllm:cache_config',
-            documentation='information of cache_config')
 
     def _unregister_vllm_metrics(self) -> None:
         for collector in list(prometheus_client.REGISTRY._collector_to_names):
@@ -180,9 +188,6 @@ class Metrics:
                 prometheus_client.REGISTRY.unregister(collector)
 
 
-# end-metrics-definitions
-
-
 class _RayGaugeWrapper:
     """Wraps around ray.util.metrics.Gauge to provide same API as
     prometheus_client.Gauge"""
@@ -190,7 +195,9 @@ class _RayGaugeWrapper:
     def __init__(self,
                  name: str,
                  documentation: str = "",
-                 labelnames: Optional[List[str]] = None):
+                 labelnames: Optional[List[str]] = None,
+                 multiprocess_mode: str = ""):
+        del multiprocess_mode
         labelnames_tuple = tuple(labelnames) if labelnames else None
         self._gauge = ray_metrics.Gauge(name=name,
                                         description=documentation,
@@ -268,10 +275,6 @@ class RayMetrics(Metrics):
         # No-op on purpose
         pass
 
-    def _create_info_cache_config(self) -> None:
-        # No-op on purpose
-        pass
-
 
 def build_1_2_5_buckets(max_value: int) -> List[int]:
     """
@@ -295,46 +298,6 @@ def build_1_2_5_buckets(max_value: int) -> List[int]:
         exponent += 1
 
 
-@dataclass
-class Stats:
-    """Created by LLMEngine for use by StatLogger."""
-    now: float
-
-    # System stats (should have _sys suffix)
-    #   Scheduler State
-    num_running_sys: int
-    num_waiting_sys: int
-    num_swapped_sys: int
-    #   KV Cache Usage in %
-    gpu_cache_usage_sys: float
-    cpu_cache_usage_sys: float
-
-    # Iteration stats (should have _iter suffix)
-    num_prompt_tokens_iter: int
-    num_generation_tokens_iter: int
-    time_to_first_tokens_iter: List[float]
-    time_per_output_tokens_iter: List[float]
-    num_preemption_iter: int
-
-    # Request stats (should have _requests suffix)
-    #   Latency
-    time_e2e_requests: List[float]
-    #   Metadata
-    num_prompt_tokens_requests: List[int]
-    num_generation_tokens_requests: List[int]
-    best_of_requests: List[int]
-    n_requests: List[int]
-    finished_reason_requests: List[str]
-
-    spec_decode_metrics: Optional["SpecDecodeWorkerMetrics"] = None
-
-
-class SupportsMetricsInfo(Protocol):
-
-    def metrics_info(self) -> Dict[str, str]:
-        ...
-
-
 def local_interval_elapsed(now: float, last_log: float,
                            local_interval: float) -> bool:
     elapsed_time = now - last_log
@@ -346,38 +309,9 @@ def get_throughput(tracked_stats: List[int], now: float,
     return float(np.sum(tracked_stats) / (now - last_log))
 
 
-class StatLoggerBase(ABC):
-    """Base class for StatLogger."""
-
-    def __init__(self, local_interval: float) -> None:
-        # Tracked stats over current local logging interval.
-        self.num_prompt_tokens: List[int] = []
-        self.num_generation_tokens: List[int] = []
-        self.last_local_log = time.time()
-        self.local_interval = local_interval
-        self.spec_decode_metrics: Optional["SpecDecodeWorkerMetrics"] = None
-
-    @abstractmethod
-    def info(self, type: str, obj: SupportsMetricsInfo) -> None:
-        raise NotImplementedError
-
-    @abstractmethod
-    def log(self, stats: Stats) -> None:
-        raise NotImplementedError
-
-    def maybe_update_spec_decode_metrics(self, stats: Stats):
-        """Save spec decode metrics (since they are unlikely
-        to be emitted at same time as log interval)."""
-        if stats.spec_decode_metrics is not None:
-            self.spec_decode_metrics = stats.spec_decode_metrics
-
-
 class LoggingStatLogger(StatLoggerBase):
     """LoggingStatLogger is used in LLMEngine to log to Stdout."""
 
-    def info(self, type: str, obj: SupportsMetricsInfo) -> None:
-        raise NotImplementedError
-
     def log(self, stats: Stats) -> None:
         """Called by LLMEngine.
            Logs to Stdout every self.local_interval seconds."""
@@ -440,10 +374,14 @@ class LoggingStatLogger(StatLoggerBase):
                 f"Number of draft tokens: {metrics.draft_tokens}, "
                 f"Number of emitted tokens: {metrics.emitted_tokens}.")
 
+    def info(self, type: str, obj: SupportsMetricsInfo) -> None:
+        raise NotImplementedError
+
 
 class PrometheusStatLogger(StatLoggerBase):
     """PrometheusStatLogger is used LLMEngine to log to Promethus."""
     _metrics_cls = Metrics
+    _gauge_cls = prometheus_client.Gauge
 
     def __init__(self, local_interval: float, labels: Dict[str, str],
                  max_model_len: int) -> None:
@@ -453,10 +391,6 @@ class PrometheusStatLogger(StatLoggerBase):
         self.metrics = self._metrics_cls(labelnames=list(labels.keys()),
                                          max_model_len=max_model_len)
 
-    def info(self, type: str, obj: SupportsMetricsInfo) -> None:
-        if type == "cache_config":
-            self.metrics.info_cache_config.info(obj.metrics_info())
-
     def _log_gauge(self, gauge, data: Union[int, float]) -> None:
         # Convenience function for logging to gauge.
         gauge.labels(**self.labels).set(data)
@@ -586,6 +520,19 @@ class PrometheusStatLogger(StatLoggerBase):
             self.last_local_log = stats.now
             self.spec_decode_metrics = None
 
+    def info(self, type: str, obj: SupportsMetricsInfo) -> None:
+        # Info type metrics are syntactic sugar for a gauge permanently set to 1
+        # Since prometheus multiprocessing mode does not support Info, emulate
+        # info here with a gauge.
+        if type == "cache_config":
+            metrics_info = obj.metrics_info()
+            info_gauge = self._gauge_cls(
+                name="vllm:cache_config_info",
+                documentation="Information of the LLMEngine CacheConfig",
+                labelnames=metrics_info.keys(),
+                multiprocess_mode="mostrecent")
+            info_gauge.labels(**metrics_info).set(1)
+
 
 class RayPrometheusStatLogger(PrometheusStatLogger):
     """RayPrometheusStatLogger uses Ray metrics instead."""

vllm/engine/metrics_types.py

@@ -0,0 +1,85 @@
+"""
+These types are defined in this file to avoid importing vllm.engine.metrics
+and therefore importing prometheus_client.
+
+This is required due to usage of Prometheus multiprocess mode to enable 
+metrics after splitting out the uvicorn process from the engine process.
+
+Prometheus multiprocess mode requires setting PROMETHEUS_MULTIPROC_DIR
+before prometheus_client is imported. Typically, this is done by setting
+the env variable before launch, but since we are a library, we need to
+do this in Python code and lazily import prometheus_client.
+"""
+
+import time
+from abc import ABC, abstractmethod
+from dataclasses import dataclass
+from typing import Dict, List, Optional, Protocol
+
+from vllm.spec_decode.metrics import SpecDecodeWorkerMetrics
+
+
+@dataclass
+class Stats:
+    """Created by LLMEngine for use by StatLogger."""
+    now: float
+
+    # System stats (should have _sys suffix)
+    #   Scheduler State
+    num_running_sys: int
+    num_waiting_sys: int
+    num_swapped_sys: int
+    #   KV Cache Usage in %
+    gpu_cache_usage_sys: float
+    cpu_cache_usage_sys: float
+
+    # Iteration stats (should have _iter suffix)
+    num_prompt_tokens_iter: int
+    num_generation_tokens_iter: int
+    time_to_first_tokens_iter: List[float]
+    time_per_output_tokens_iter: List[float]
+    num_preemption_iter: int
+
+    # Request stats (should have _requests suffix)
+    #   Latency
+    time_e2e_requests: List[float]
+    #   Metadata
+    num_prompt_tokens_requests: List[int]
+    num_generation_tokens_requests: List[int]
+    best_of_requests: List[int]
+    n_requests: List[int]
+    finished_reason_requests: List[str]
+
+    spec_decode_metrics: Optional["SpecDecodeWorkerMetrics"] = None
+
+
+class SupportsMetricsInfo(Protocol):
+
+    def metrics_info(self) -> Dict[str, str]:
+        ...
+
+
+class StatLoggerBase(ABC):
+    """Base class for StatLogger."""
+
+    def __init__(self, local_interval: float) -> None:
+        # Tracked stats over current local logging interval.
+        self.num_prompt_tokens: List[int] = []
+        self.num_generation_tokens: List[int] = []
+        self.last_local_log = time.time()
+        self.local_interval = local_interval
+        self.spec_decode_metrics: Optional["SpecDecodeWorkerMetrics"] = None
+
+    @abstractmethod
+    def log(self, stats: Stats) -> None:
+        raise NotImplementedError
+
+    @abstractmethod
+    def info(self, type: str, obj: SupportsMetricsInfo) -> None:
+        raise NotImplementedError
+
+    def maybe_update_spec_decode_metrics(self, stats: Stats):
+        """Save spec decode metrics (since they are unlikely
+        to be emitted at same time as log interval)."""
+        if stats.spec_decode_metrics is not None:
+            self.spec_decode_metrics = stats.spec_decode_metrics

vllm/entrypoints/openai/api_server.py

@@ -2,7 +2,9 @@ import asyncio
 import importlib
 import inspect
 import multiprocessing
+import os
 import re
+import tempfile
 from argparse import Namespace
 from contextlib import asynccontextmanager
 from http import HTTPStatus
@@ -12,7 +14,6 @@ from fastapi import APIRouter, FastAPI, Request
 from fastapi.exceptions import RequestValidationError
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import JSONResponse, Response, StreamingResponse
-from prometheus_client import make_asgi_app
 from starlette.routing import Mount
 
 import vllm.envs as envs
@@ -54,6 +55,7 @@ openai_serving_chat: OpenAIServingChat
 openai_serving_completion: OpenAIServingCompletion
 openai_serving_embedding: OpenAIServingEmbedding
 openai_serving_tokenization: OpenAIServingTokenization
+prometheus_multiproc_dir: tempfile.TemporaryDirectory
 
 logger = init_logger('vllm.entrypoints.openai.api_server')
 
@@ -109,6 +111,21 @@ async def build_async_engine_client(args) -> AsyncIterator[AsyncEngineClient]:
 
     # Otherwise, use the multiprocessing AsyncLLMEngine.
     else:
+        if "PROMETHEUS_MULTIPROC_DIR" not in os.environ:
+            # Make TemporaryDirectory for prometheus multiprocessing
+            # Note: global TemporaryDirectory will be automatically
+            #   cleaned up upon exit.
+            global prometheus_multiproc_dir
+            prometheus_multiproc_dir = tempfile.TemporaryDirectory()
+            os.environ[
+                "PROMETHEUS_MULTIPROC_DIR"] = prometheus_multiproc_dir.name
+        else:
+            logger.warning(
+                "Found PROMETHEUS_MULTIPROC_DIR was set by user. "
+                "This directory must be wiped between vLLM runs or "
+                "you will find inaccurate metrics. Unset the variable "
+                "and vLLM will properly handle cleanup.")
+
         # Select random path for IPC.
         rpc_path = get_open_zmq_ipc_path()
         logger.info("Multiprocessing frontend to use %s for RPC Path.",
@@ -149,13 +166,38 @@ async def build_async_engine_client(args) -> AsyncIterator[AsyncEngineClient]:
             # Wait for server process to join
             rpc_server_process.join()
 
+            # Lazy import for prometheus multiprocessing.
+            # We need to set PROMETHEUS_MULTIPROC_DIR environment variable
+            # before prometheus_client is imported.
+            # See https://prometheus.github.io/client_python/multiprocess/
+            from prometheus_client import multiprocess
+            multiprocess.mark_process_dead(rpc_server_process.pid)
+
 
 router = APIRouter()
 
 
 def mount_metrics(app: FastAPI):
-    # Add prometheus asgi middleware to route /metrics requests
+    # Lazy import for prometheus multiprocessing.
-    metrics_route = Mount("/metrics", make_asgi_app())
+    # We need to set PROMETHEUS_MULTIPROC_DIR environment variable
+    # before prometheus_client is imported.
+    # See https://prometheus.github.io/client_python/multiprocess/
+    from prometheus_client import (CollectorRegistry, make_asgi_app,
+                                   multiprocess)
+
+    prometheus_multiproc_dir_path = os.getenv("PROMETHEUS_MULTIPROC_DIR", None)
+    if prometheus_multiproc_dir_path is not None:
+        logger.info("vLLM to use %s as PROMETHEUS_MULTIPROC_DIR",
+                    prometheus_multiproc_dir_path)
+        registry = CollectorRegistry()
+        multiprocess.MultiProcessCollector(registry)
+
+        # Add prometheus asgi middleware to route /metrics requests
+        metrics_route = Mount("/metrics", make_asgi_app(registry=registry))
+    else:
+        # Add prometheus asgi middleware to route /metrics requests
+        metrics_route = Mount("/metrics", make_asgi_app())
+
     # Workaround for 307 Redirect for /metrics
     metrics_route.path_regex = re.compile('^/metrics(?P<path>.*)$')
     app.routes.append(metrics_route)