Statsd host metrics (python psutil)

agent.py

#!/usr/bin/env python
# coding: utf-8

import sys
reload(sys)
sys.setdefaultencoding('utf-8')

import os
import time

import psutil
import statsd


def send(stats):
    t1 = time.time()
    print('CPU:\n====')
    cpu_percent = psutil.cpu_percent(interval=1)
    stats.gauge('cpu.percent.all', cpu_percent)
    print('cpu.percent.all: {}'.format(cpu_percent))
    m1, m5, m15 = os.getloadavg()
    stats.gauge('cpu.load.1', m1)
    stats.gauge('cpu.load.5', m5)
    stats.gauge('cpu.load.15', m15)
    print('cpu.load.1: {}'.format(m1))
    print('cpu.load.5: {}'.format(m5))
    print('cpu.load.15: {}'.format(m15))

    print(u'\n内存:\n=====')
    vmem = psutil.virtual_memory()
    for k in ['total', 'used', 'buffers', 'cached']:
        key = 'memory.virtual.{}'.format(k)
        value = getattr(vmem, k)
        stats.gauge(key, value)
        print('{}: {}'.format(key, value))
    smem = psutil.swap_memory()
    stats.gauge('memory.swap.used', smem.used)
    print('memory.swap.used: {}'.format(smem.used))

    print(u'\n磁盘:\n=====')
    disk_parts = psutil.disk_partitions()
    for disk_part in disk_parts:
        mountpoint = disk_part.mountpoint.replace('/', '_')
        try:
            disk_usage = psutil.disk_usage(disk_part.mountpoint)
            stats.gauge('disk.usage.percent.{}'.format(mountpoint), disk_usage.percent)
            stats.gauge('disk.usage.total.{}'.format(mountpoint), disk_usage.total)
            print('disk.usage.percent.{}: {}%'.format(mountpoint, disk_usage.percent))
            print('disk.usage.total.{}: {}'.format(mountpoint, disk_usage.total))
        except OSError:
            pass
    disk_io = psutil.disk_io_counters(perdisk=False)
    for k in ['read_count', 'write_count', 'read_bytes', 'write_bytes']:
        key = 'disk.io.{}.all'.format(k)
        value = getattr(disk_io, k)
        stats.gauge(key, value)
        print('{}: {}'.format(key, value))

    print(u'\n网络:\n====')
    net_io = psutil.net_io_counters(pernic=False)
    stats.gauge('net.io.bytes_sent.all', net_io.bytes_sent)
    stats.gauge('net.io.bytes_recv.all', net_io.bytes_recv)
    print('net.io.bytes_sent.all: {}'.format(net_io.bytes_sent))
    print('net.io.bytes_recv.all: {}'.format(net_io.bytes_recv))    

    print(u'\n其它:\n=====')
    uptime = int(time.time() - psutil.boot_time())
    stats.gauge('uptime', uptime)
    h = uptime / 3600
    m = (uptime % 3600) / 60
    s = uptime % 60
    print('uptime: {} hour, {} minutes, {} seconds'.format(h, m, s))
    print('[Cost]: {}s'.format(time.time() - t1))


def main():
    stats = statsd.client.StatsClient()
    while True:
        send(stats)
        time.sleep(15)


if __name__ == '__main__':
    main()    

metrics.md

CPU

# [占用百分比]
 psutil.cpu_percent()
# [Statsd key]
cpu.percent.all

# [负载]: Return the number of processes in the system run queue averaged over the last 1, 5, and 15 minutes
 os.getloadavg()
# [Statsd key]
cpu.load.1    # 1 minutes
cpu.load.5    # 5 minutes
cpu.load.15   # 15 minutes

内存

# [虚拟内存]
 psutil.virtual_memory().{ total, used, buffers, cached }
# [Statsd key]
memory.virtual.total     # ??? 是常量
memory.virtual.used
memory.virtual.buffers
memory.virtual.cached

# [Swap]
psutil.swap_memory().used
# [Statsd key]
memory.swap.used

磁盘

# [分区]
psutil.disk_partitions()

# [某个分区的使用量]  ??? total 是常量
psutil.disk_usage(path)
# [Statsd key]
disk.usage.percent._       # mountpoint = /
disk.usage.total._         # mountpoint = /
disk.usage.percent._boot   # mountpoint = /boot
disk.usage.total._boot     # mountpoint = /boot

# [使用率(总体的)]
 psutil.disk_io_counters(perdisk=False).{ read_count, write_count, read_bytes, write_bytes }
# [Statsd key]
disk.io.read_count.all
disk.io.write_count.all
disk.io.read_bytes.all
disk.io.write_bytes.all

网络

# [使用率(总体)]
 psutil.net_io_counters(pernic=False).{ bytes_sent, bytes_recv }
# [Statsd key]
net.io.bytes_sent.all
net.io.bytes_recv.all

其它

# [开机时间]
int(time.time() - psutil.boot_time())
# [Statsd key]
uptime