Linux shell 脚本监控cpu,内存,硬盘,网络,是否存活
Posted
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了Linux shell 脚本监控cpu,内存,硬盘,网络,是否存活相关的知识,希望对你有一定的参考价值。
写脚本的背景:脚本实现简单的监控,而不需要用非常的重的监控软件完成。
脚本如下
#!/bin/sh # 脚本放入到/usr/local/bin # chmod 755 check_server.sh # crontab 中添加 source /etc/bashrc #------监控阈值 DISK_space_warn=90 CPU_load_warn=5 CPU_use_warn=50 MEM_use_warn=95 #SWAP_use_warn=50 Net_SYN_count_warn=200 #判断参数 if [ $# -ne 2 ]; then hint_msg="usage: $0 Monitoring_type,Monitoring_type phone,phone,phone" echo $hint_msg echo "sh $0 disk,cpu,mem,net,io,alive" exit -1 fi #------监控项 Monitoring_type_tmp=$1 Monitoring_type=$(echo $Monitoring_type_tmp | sed -e ‘s/,/ /g‘) now=`date -u -d"+8 hour" +‘%Y-%m-%d %H:%M:%S‘` #------机器操作系统 OS_version=‘‘ if grep -q ‘CentOS release 6‘ /etc/redhat-release ; then OS_version=‘CentOS6‘ else OS_version=‘CentOS7‘ fi #------ip地址 localip=`hostname -I | tr ‘ ‘ ‘\n‘ | grep -E ‘(^10\.|^172\.(1[6-9]|2[0-9]|31)|^192\.168)‘ | head -n 1` send_warning() { 发短信的函数 发送下面的msg变量 } #------监控CPU相关信息 function sub_cpu(){ cpu_num=`grep -c ‘model name‘ /proc/cpuinfo` #------cpu 负载 15 minutes load_15=`cat /proc/loadavg | awk ‘{print $2}‘` average_load=`echo "scale=2;a=$load_15/$cpu_num;if(length(a)==scale(a)) print 0;print a" | bc` average_int=`echo $average_load | cut -f 1 -d "."` if [ "${average_int}" -ge "${CPU_load_warn}" ];then msg=${HOSTNAME}" "${localip}" System load average of 15 minutes is "${average_load}" more than "${CPU_load_warn} echo "${now} ${msg}" >> log send_warning fi #------cpu 使用率 if [ ${OS_version} == ‘CentOS6‘ ];then cpu_idle=`top -b -n 1 | grep Cpu | awk ‘{print $5}‘ | cut -f 1 -d "."` else cpu_idle=`top -b -n 1 | grep Cpu | awk ‘{print $8}‘ | cut -f 1 -d "."` fi CPU_use=`expr 100 - $cpu_idle` if [ "${CPU_use}" -ge "${CPU_use_warn}" ];then msg=${HOSTNAME}" "${localip}" CPU utilization is "${CPU_use}"% more than "${CPU_use_warn}"%" echo "${now} ${msg}" >> log send_warning fi } #------监控内存 function sub_mem(){ MEM_use=`free | grep "Mem" | awk ‘{printf("%d", $3*100/$2)}‘` if [ $MEM_use -ge $MEM_use_warn ];then msg=${HOSTNAME}" "${localip}" Mem_used:"${MEM_use}"% more than "${MEM_use_warn}"%" echo "${now} ${msg}" >> log send_warning fi #SWAP有很多环境为0 例如在AWS云,腾讯云上 #SWAP_use=`free | grep "Swap" | awk ‘{printf("%d", $3*100/$2)}‘` #if [ $SWAP_use -ge $SWAP_use_warn ];then # # send_warning #fi } #------监控硬盘空间 function sub_disk(){ #[注意这里用df -P] for DISK_space in `df -P | grep /dev | grep -v -E ‘(tmp|boot)‘ | awk ‘{print $5}‘ | cut -f 1 -d "%" ` do if [ $DISK_space -ge "${DISK_space_warn}" ]; then msg=${HOSTNAME}" "${localip}" Hard disk space :"${DISK_space}"% more than "${DISK_space_warn}"%" echo "${now} ${msg}" >> log send_warning fi done } #------监控网络相关 function sub_net(){ #------syn 半连接数 Net_SYN_count=`ss -an | grep -ic syn` if [ "${Net_SYN_count}" -ge "${Net_SYN_count_warn}" ]; then msg=${HOSTNAME}" "${localip}" Net SYN count :"${Net_SYN_count}" more than "${Net_SYN_count_warn} echo "${now} ${msg}" >> log send_warning fi } #------监控项 for type in $Monitoring_type do [ "${type}" == ‘disk‘ ] && sub_disk [ "${type}" == ‘cpu‘ ] && sub_cpu [ "${type}" == ‘mem‘ ] && sub_mem [ "${type}" == ‘net‘ ] && sub_net [ "${type}" == ‘io‘ ] && sub_io #[ "${type}" == ‘alive‘ ] && sub_alive done #判断端口是否通 也是判断是否alive的 #!/bin/bash #Alive=`echo -e "\n" | telnet 192.168.1.30 22 | grep Connected | wc -l` #if [ "$Alive" == 1 ];then # echo "0" #如果JG等于1,端口为通,输出0 #else # echo "1" #如果JG等于0,端口不通,输出1 #fi
进行测试
本文出自 “走起运维” 博客,请务必保留此出处http://yuanji6699.blog.51cto.com/11568362/1910172
以上是关于Linux shell 脚本监控cpu,内存,硬盘,网络,是否存活的主要内容,如果未能解决你的问题,请参考以下文章
找一找青春Linux 性能监控之CPU&内存&IO监控Shell脚本1[转]