linux下对服务器性能监控shell脚本

 #!/bin/bash

 #提取本服务器的IP地址信息
 ENO1=`ifconfig | sed -n '1,1p' | awk -F ' ' '{print $1}'`
 IP=`ifconfig ${ENO1}| grep "inet addr" | cut -f  -d ":" | cut -f  -d " "`
 echo 'eno: '${ENO1}
 echo 'IP: '${IP}

 # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
 # 各个监控警告值
 WARN_LOAD=1.1
 WARN_CPU=0.95
 WARN_MEM=0.93
 # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #

 # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
 CPU=0.0
 LOAD1=0.0
 LOAD5=0.0
 LOAD15=0.0
 MEM=0.0
 # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #

 send_mail() {
     echo 'send mail '  "${1}" "${2}"
 }

 # 、监控系统负载的变化情况.
 system_load_func()
 {
     cpu_num=`grep -c 'model name' /proc/cpuinfo`

     load=$(uptime | awk -F 'load average: ' '{print $2}')
     load_1=$(echo $load | awk -F ', ' '{print $1}')
     load_5=$(echo $load | awk -F ', ' '{print $2}')
     load_15=$(echo $load | awk -F ', ' '{print $3}')

     #计算当前系统单个核心平均负载值，结果小于1.0时前面个位数补0。
     LOAD1=`echo "scale=2;a=${load_1}/${cpu_num};if(length(a)==scale(a)) print 0;print a" | bc`
     LOAD5=`echo "scale=2;a=${load_5}/${cpu_num};if(length(a)==scale(a)) print 0;print a" | bc`
     LOAD15=`echo "scale=2;a=${load_15}/${cpu_num};if(length(a)==scale(a)) print 0;print a" | bc`

     if [ `echo "${LOAD15} > ${WARN_LOAD}" | bc` -eq  ]
     then
         send_mail "${IP}服务器15分钟的系统平均负载为${LOAD15}, 超过警戒值${WARN_LOAD}, 请立即处理!" "$IP 服务器系统负载告警!"
     fi
 }

 # 、监控系统cpu的情况.
 cpu_func()
 {
     #取当前空闲cpu百份比值(只取整数部分)
     cpu_idle=`top -b -d 0.1 -n  | grep Cpu | tail -n  | awk '{print $8}' | cut -f  -d "."`
     CPU=`echo "scale=2; (100 - ${cpu_idle}) / 100" | bc`

     if [ `echo "${CPU} > ${WARN_CPU}" | bc` -eq  ]
     then
         send_mail "${IP}服务器cpu使用${CPU},请及时处理." "$IP 服务器CPU告警"
     fi
 }

 #. 监控物理内存
 mem_func()
 {
     mem=$(free -m | sed -n '2,2p')
     mem_total=`echo ${mem} | awk -F ' ' '{print $2}'`
     mem_used=`echo ${mem} | awk -F ' ' '{print $3}'`
     # mem_free=`echo ${mem} | awk -F ' ' '{print $4}'`

     if [ $mem_used -ne  ]
     then
         MEM=`echo "scale=2; ${mem_used} / ${mem_total}" | bc`
         if [ `echo "${MEM} > ${WARN_MEM}" | bc` -eq  ]
         then
             send_mail "$IP服务器物理内存已使用 ${MEM},请及时处理." "$IP 服务器内存告警"
         fi
     fi
 }

 CNT=
 print_info() {
     echo "# # # # # # # # # # # [${CNT}] # # # # # # # # # # # # # #"
     echo cpu: ${CPU}
     echo load1: ${LOAD1}
     echo load5: ${LOAD5}
     echo load15: ${LOAD15}
     echo mem: ${MEM}
     echo "# # # # # # # # # # # # # # # # # # # # # # # # # # #"
     (( CNT +=  ))
     # exit
 }

 while true
 do
     system_load_func
     cpu_func
     mem_func
     print_info
     sleep
 done

注: 需要下载bc命令:

sudo apt-get install bc

这里我只在脚本里面写了获取各个监控信息的数据, 发送邮件和数据写入数据库这些我省略掉了. 发送邮件我使用的是python, 通过该shell脚本调用python发送邮件.(因为用linux自带的我发现会出现很多垃圾邮件, 怎么解决我没有去研究). 得到的数据我会写入数据库记录, 在这里为了通用性, 我把写入数据库部分省略掉了.