批量监控关键业务站点nginx的http状态码

一些业务站点比较关键，比如API接口或者web站点，需要对出现的访问http状态码进行监控，比如监控10分钟以内，或者最近的1000次访问，当出现的50x状态码过多时触发告警，可以让一些问题暴露出来及时处理

1.编写脚本

# cat /usr/local/zabbix_agents_3.2.0/scripts/web_nginx_code.sh 
#!/bin/bash 
# function:monitor store nginx access error code

web_domain_discovery () { 
WEB_DOMAIN=($(cat  /usr/local/zabbix_agents_3.2.0/scripts/web_site.txt|grep -v "^#")) 
        printf '{
' 
        printf '	"data":[
' 
for((i=0;i<${#WEB_DOMAIN[@]};++i)) 
    { 
    num=$(echo $((${#WEB_DOMAIN[@]}-1))) 
        if [ "$i" != ${num} ]; 
            then 
        printf "		{ 
" 
        printf "			"{#SITENAME}":"${WEB_DOMAIN[$i]}"},
" 
            else 
                printf  "		{ 
" 
                printf  "			"{#SITENAME}":"${WEB_DOMAIN[$num]}"}]}
" 
        fi 
    } 
} 

# 统计nginx access log中的50x代码个数
error_code_count () { 
    web_site=$1
    if [ ${web_site} == 'store.chinasoft.jp' ];then
        #/usr/bin/cat /data/www/logs/nginx_log/access/${web_site}.access.log |awk '{print $1" "$10" "$11}'|grep 'HTTP/1.1" 50'|uniq |wc -l
        tail -n 1000 /data/www/logs/nginx_log/access/${web_site}.access.log |awk '{print $1" "$10" "$11}'|grep 'HTTP/1.1" 50'|uniq |wc -l
        exit 0
    fi
    #/usr/bin/cat /data/www/logs/nginx_log/access/${web_site}_access.log |awk '{print $1" "$10" "$11}'|grep 'HTTP/1.1" 50'|uniq |wc -l
    tail -n 1000 /data/www/logs/nginx_log/access/${web_site}_access.log |awk '{print $1" "$10" "$11}'|grep 'HTTP/1.1" 50'|uniq |wc -l
} 

last10_mins_error_code_count () { 
    web_site=$1
    if [ ${web_site} == 'store.chinasoft.jp' ];then
        /usr/bin/cat /data/www/logs/nginx_log/access/${web_site}.access.log |awk '{print $1" "$10" "$11}'|grep 'HTTP/1.1" 50'|uniq |wc -l
        exit 0
    fi
    #/usr/bin/cat /data/www/logs/nginx_log/access/${web_site}_access.log |awk '{print $1" "$10" "$11}'|grep 'HTTP/1.1" 50'|uniq |wc -l
    # 统计最近 10 分钟的 50x 错误
    /usr/bin/tac /data/www/logs/nginx_log/access/${web_site}_access.log| awk 'BEGIN{ "date -d "-10 minute" +"%H:%M:%S"" | getline min10ago } { if (substr($4, 14) > min10ago) print $0;else exit }' | tac| awk '{print $1" "$10" "$11}' | grep 'HTTP/1.1" 50'|uniq|wc -l
} 

 
case "$1" in
    web_domain_discovery) 
        web_domain_discovery 
    ;;
    error_code_count) 
        error_code_count $2
    ;; 
        last10_mins_error_code_count)
                last10_mins_error_code_count $2
        ;;
*)
 
echo "Usage:$0 {web_web_discovery|error_code_count|last10_mins_error_code_count}" 
;; 
esac

站点的读取文件格式

# cat /usr/local/zabbix_agents_3.2.0/scripts/web_site.txt 
account.chinasoft.cn
distriapi.chinasoft.cn
innerapi.chinasoft.cn
masterapi.chinasoft.cn
mm.chinasoft.cn
userapi.chinasoft.cn

2.编写监控配置

# cat /usr/local/zabbix_agents_3.2.0/conf/zabbix_agentd/web_nginx_code_discovery.conf 
UserParameter=web.domain.discovery,/usr/local/zabbix_agents_3.2.0/scripts/web_nginx_code.sh web_domain_discovery
UserParameter=web.domain.code[*],/usr/local/zabbix_agents_3.2.0/scripts/web_nginx_code.sh error_code_count $1
UserParameter=web.domain.10mins.code[*],/usr/local/zabbix_agents_3.2.0/scripts/web_nginx_code.sh last10_mins_error_code_count $1

3.创建监控项

name和key都是：web.domain.discovery

监控项

10分钟内的状态码

name: web.domain.10mins.code ON $1

key: web.domain.10mins.code[{#SITENAME},]

监控项

name:web.domain.code ON $1

key: web.domain.code[{#SITENAME},]

触发器

name: {#SITENAME} last 10 minutes nginx 50x great than 10

表达式

{Template alisz nginx site access error_code count:web.domain.10mins.code[{#SITENAME},].last()}>50

10分钟超过50个50x就报警

最近的1000个访问超过200触发报警

name:{#SITENAME}  nginx 50x code great than 20%

触发器：

{Template alisz nginx site access error_code count:web.domain.code[{#SITENAME},].last()}>200