通过 python 获取内存使用情况

Question

我在使用 psutil 检索系统内存使用时遇到问题。

当我运行我的代码时，它会正确检索 cpu 用法，但是当它尝试获取内存时，它会抛出 AccessDenied 错误。

代码：

if(pi['name']=='java.exe'):
    pro=psutil.Process(0)
    cpu=pro.cpu_percent(1)
    memory=pro.memory_full_info().uss/(1024*1024)
    return memory,cpu

错误信息：

psutil.AccessDenied (pid=0)

如何获取内存使用情况？

Answer 1

psutil.Process() 接受一个参数——进程标识符 (PID)。如果调用时不带参数或使用 None，则使用当前进程的 pid。

在 Window 上，pid 0 是 系统空闲进程，它需要 SYSTEM 权限。

pro=psutil.Process(0) # here's your problem

如果您想获得 java.exe 的 memory/cpu 用法，您必须获得 pid，然后使用 psutil。在下面的示例中，我使用了 firefox.exe:

import psutil
from ctypes import *
from ctypes.wintypes import *

class PROCESSENTRY32W(Structure):
    _fields_ = [("dwSize", DWORD),
                ("cntUsage", DWORD),
                ("th32ProcessID", DWORD),
                ("th32DefaultHeapID", POINTER(ULONG)),
                ("th32ModuleID", DWORD),
                ("cntThreads", DWORD),
                ("th32ParentProcessID", DWORD),
                ("pcPriClassBase", LONG),
                ("dwFlags", DWORD),
                ("szExeFile", c_wchar * MAX_PATH)]

def get_processes():
    TH32CS_SNAPPROCESS = 0x00000002
    INVALID_HANDLE_VALUE = -1

    CreateToolhelp32Snapshot = windll.kernel32.CreateToolhelp32Snapshot
    CreateToolhelp32Snapshot.argtypes = [DWORD, DWORD]
    CreateToolhelp32Snapshot.restype = HANDLE

    Process32FirstW = windll.kernel32.Process32FirstW
    Process32FirstW.argtypes = [HANDLE, POINTER(PROCESSENTRY32W)]
    Process32FirstW.restype = BOOL

    Process32NextW = windll.kernel32.Process32NextW
    Process32NextW.argtypes = [HANDLE, POINTER(PROCESSENTRY32W)]
    Process32NextW.restype = BOOL

    CloseHandle = windll.kernel32.CloseHandle
    CloseHandle.argtypes = [HANDLE]
    CloseHandle.restype = BOOL

    snapshot = CreateToolhelp32Snapshot(TH32CS_SNAPPROCESS, 0)
    if snapshot == INVALID_HANDLE_VALUE:
        return []

    try:
        process_entry = PROCESSENTRY32W()
        process_entry.dwSize = sizeof(PROCESSENTRY32W)
        if not Process32FirstW(snapshot, byref(process_entry)):
            return []

        processes = []
        while Process32NextW(snapshot, byref(process_entry)):
            processes.append((process_entry.th32ProcessID, process_entry.szExeFile))
    finally:
        CloseHandle(snapshot)
        #print("Closed handle")

    return processes

for pid, name in get_processes():
    if name == "firefox.exe":
        pro = psutil.Process(pid)
        cpu = pro.cpu_percent(1.0)
        memory = pro.memory_full_info().uss/(1024*1024)
        print(memory, cpu)

输出：

471.2890625 1.6
127.15234375 25.0
18.50390625 0.0
668.09375 0.0
706.6875 0.0
512.671875 0.0
648.953125 4.7
1037.0859375 4.7
1160.1015625 98.4
337.46484375 0.0
1418.72265625 1.6
0.90625 0.0
0.90625 0.0

通过 python 获取内存使用情况

Getting memory usage by python

python

memory

cpu

process