Python监控服务器利器--psutil

服务器的监控通过安装一些常用的监控软件之外，有时也需要运行一些shell或Python脚本；shell下可以使用系统自带的ps/free/top/df等shell命令，Python可以调用subprocess等模块来运行shell命令，不过这么做就比较麻烦。这里有一个比较好用的第三方模块：psutil。

psutil是一个跨平台的库，用于在Python中检索有关运行进程和系统利用率（CPU，内存，磁盘，网络，传感器）的信息。它主要用于系统监视，分析，限制进程资源和运行进程的管理。它实现了UNIX命令行工具提供的许多功能，例如：ps，top，lsof，netstat，ifconfig，who，df，kill，free，nice，ionice，iostat，iotop，uptime，pidof，tty，taskset，pmap。 psutil目前支持以下平台：

Linux 
Windows
OSX,
FreeBSD, OpenBSD, NetBSD
Sun Solaris
AIX
...

等装有Python2.6至3.6的32-bit和64-bit架构. 也可以在PyPy上运行。
安装
文中示例均在Python版本3.6环境下运行；
# pip3 install psutil
常用模块
获取psutil版本信息

In [1]: import psutil
In [2]: psutil.version_info
Out[2]: (5, 4, 3

获取CPU信息

In [3]: psutil.cpu_count() # 逻辑CPU核数
Out[3]: 4
In [4]: psutil.cpu_count(logical=False) # 物理CPU核数
Out[4]: 2
In [5]: psutil.cpu_times() # CPU的用户、系统、空闲时间
Out[5]: scputimes(user=240773.0, nice=0.0, system=96416.32, idle=1161930.41)
In [9]: psutil.cpu_percent(percpu=True) # 获取每个CPU的使用率，类似TOP命令
Out[9]: [43.3, 22.0, 42.0, 23.0]
In [10]: top = [psutil.cpu_percent(interval=i, percpu=True) for i in range(10)] #设置每秒刷新时间间隔，统计十次的结果
In [11]: top
Out[11]:
[[40.8, 19.7, 38.5, 20.7],
[25.7, 5.9, 13.0, 5.0],
[35.0, 15.6, 30.0, 14.4],
[23.7, 7.0, 18.3, 7.4],
[38.5, 17.0, 34.2, 17.5],
[37.2, 19.6, 36.3, 20.0],
[29.6, 16.6, 28.8, 16.8],
[37.7, 19.0, 35.4, 18.7],
[30.8, 16.3, 26.9, 16.5],
[44.2, 27.9, 41.5, 28.6]]

获取内存信息

In [13]: psutil.virtual_memory() #获取内存统计数据，单位bytes，我这里8G内存
Out[13]: svmem(total=8589934592, available=1891045376, percent=78.0, used=6053986304, free=15130624, active=1878392832, inactive=1875914752, wired=2299678720)
In [14]: psutil.swap_memory() # 获取swap的统计数据
Out[14]: sswap(total=2147483648, used=1340866560, free=806617088, percent=62.4, sin=126090076160, sout=3524710400)

获取磁盘信息

In [17]: psutil.disk_partitions() #获取磁盘分区信息
Out[17]: [sdiskpart(device='/dev/disk1', mountpoint='/', fstype='hfs', opts='rw,local,rootfs,dovolfs,journaled,multilabel')]
In [20]: psutil.disk_usage('/') # 获取分区使用情况，这里使用了25.4%
Out[20]: sdiskusage(total=499055067136, used=126482944000, free=372309979136, percent=25.4)
In [22]: psutil.disk_io_counters() #磁盘IO情况
Out[22]: sdiskio(read_count=7364142, write_count=6510641, read_bytes=282106464256, write_bytes=261763244544, read_time=2608778, write_time=1095259)

获取网络信息

In [23]: psutil.net_if_stats() # 获取网卡接口状态
Out[23]:
{'awdl0': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1484),
'bridge0': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1500),
'en0': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1500),
'en1': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_FULL: 2>, speed=0, mtu=1500),
'en2': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_FULL: 2>, speed=0, mtu=1500),
'gif0': snicstats(isup=False, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1280),
'lo0': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=16384),
'p2p0': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=2304),
'stf0': snicstats(isup=False, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1280),
'utun0': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=2000),
'utun1': snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1352)}
In [25]: psutil.net_if_stats().get("en0") #获取单个网卡en0的状态
Out[25]: snicstats(isup=True, duplex=<NicDuplex.NIC_DUPLEX_UNKNOWN: 0>, speed=0, mtu=1500)
In [26]: psutil.net_if_addrs() # 获取所有网卡的地址信息
Out[26]:
{'awdl0': [snic(family=<AddressFamily.AF_LINK: 18>, address='36:7d:f3:80:6e:4e', netmask=None, broadcast=None, ptp=None),
snic(family=<AddressFamily.AF_INET6: 30>, address='fe80::347d:f3ff:fe80:6e4e%awdl0', netmask='ffff:ffff:ffff:ffff::', broadcast=None, ptp=None)],
'bridge0': [snic(family=<AddressFamily.AF_LINK: 18>, address='4a:00:02:c0:33:70', netmask=None, broadcast=None, ptp=None)],
'en0': [snic(family=<AddressFamily.AF_INET: 2>, address='192.168.0.101', netmask='255.255.255.0', broadcast='192.168.0.255', ptp=None),
snic(family=<AddressFamily.AF_LINK: 18>, address='ac:bc:32:91:32:8b', netmask=None, broadcast=None, ptp=None),
snic(family=<AddressFamily.AF_INET6: 30>, address='fe80::1476:ce7e:210a:2e32%en0', netmask='ffff:ffff:ffff:ffff::', broadcast=None, ptp=None)],
'en1': [snic(family=<AddressFamily.AF_LINK: 18>, address='4a:00:02:c0:33:70', netmask=None, broadcast=None, ptp=None)],
'en2': [snic(family=<AddressFamily.AF_LINK: 18>, address='4a:00:02:c0:33:71', netmask=None, broadcast=None, ptp=None)],
'lo0': [snic(family=<AddressFamily.AF_INET: 2>, address='127.0.0.1', netmask='255.0.0.0', broadcast=None, ptp=None),
snic(family=<AddressFamily.AF_INET6: 30>, address='::1', netmask='ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff', broadcast=None, ptp=None),
snic(family=<AddressFamily.AF_INET6: 30>, address='fe80::1%lo0', netmask='ffff:ffff:ffff:ffff::', broadcast=None, ptp=None)],
'p2p0': [snic(family=<AddressFamily.AF_LINK: 18>, address='0e:bc:32:91:32:8b', netmask=None, broadcast=None, ptp=None)],
'utun0': [snic(family=<AddressFamily.AF_INET6: 30>, address='fe80::583c:77a0:6b93:b045%utun0', netmask='ffff:ffff:ffff:ffff::', broadcast=None, ptp=None)],
'utun1': [snic(family=<AddressFamily.AF_INET: 2>, address='10.5.200.244', netmask=None, broadcast=None, ptp='10.5.200.244')]}
# 获取en0网卡的地址， 这里包括mac和ipv6地址
In [40]: for addr in psutil.net_if_addrs().get("en0"):
...: print(addr.address)
192.168.0.101
ac:bc:32:91:32:8b
fe80::1476:ce7e:210a:2e32%en0
In [43]: psutil.net_io_counters() # 获取网络读写字节/包的个数
Out[43]: snetio(bytes_sent=174614221, bytes_recv=586279725, packets_sent=863903, packets_recv=873583, errin=0, errout=0, dropin=0, dropout=0)
In [45]: psutil.net_connections() # 获取网络连接信息，注意这里需要root权限。

获取进程信息：

In [46]: psutil.pids() # 获取所有进程ID
In [47]: psutil.Process(61) # 获取指定PID的进程信息
Out[47]: psutil.Process(pid=61, name='dsAccessService', started='2018-02-26 09:57:04')
In [49]: psutil.Process(45573).exe() # 获取进程的exe路径
Out[49]: '/usr/local/Cellar/python3/3.6.3/Frameworks/Python.framework/Versions/3.6/Resources/Python.app/Contents/MacOS/Python'
In [50]: psutil.Process(45573).name() # 获取进程名称
Out[50]: 'Python'
In [52]: psutil.Process(45573).cmdline() # 获取进程启动的命令
Out[52]:
['/usr/local/Cellar/python3/3.6.3/Frameworks/Python.framework/Versions/3.6/Resources/Python.app/Contents/MacOS/Python',
'/usr/local/bin/ptipython']
In [56]: psutil.Process(45573).num_threads() # 获取进程的线程数量
Out[56]: 3
In [57]: psutil.Process(45573).environ() # 获取进程的环境变量信息

总结：
使用psutil模块可以做到比较全面的对系统的监控，如果你正在考虑用Python做一个监控系统或者脚本工具，可以有些考虑此模块。