• aws cloudwatch监控怎么通过钉钉机器人报警


    最近在完善海外业务在aws服务的CloudWatchh监控,发现CloudWatch报警通知要通过aws的sns服务,直接支持的通道有短信和邮件,但是我们想推到钉钉群里面的群机器人里面这个就要借助aws的Lambda函数服务

    然后选择用sns来触发

    python脚本内容变量event是sns传过来的消息内容,有点坑的是CloudWatch的报警信息很乱还要稍微优化一下

    # -*- coding: utf-8 -*-
    import json
    import os
    import re
    from botocore.vendored import requests
    
    
    def size_b_to_other(size):
        """用于转换容量单位"""
        units = ['B', 'KB', 'MB', 'GB', 'TB']
        # 处理异常
        if size <= 0:
            return False
    
        # 遍历单位的位置并用取整除赋值
        for unit in units:
            if size >= 1024:
                size //= 1024
            else:
                size_h = '{} {}'.format(size, unit)
                return size_h
    
        size_h = '{} {}'.format(size, unit)
        return size_h
        
        
    def lambda_handler(event, context):
        token = os.getenv('token')
        url = "https://oapi.dingtalk.com/robot/send?access_token="
        headers = {'Content-Type': 'application/json'}
        # 解析要使用的字段
        Sns = event['Records'][0]['Sns']
        Subject = Sns['Subject']
        if "ALARM" in Subject:
            title = "======报警信息======"
        elif "OK" in Subject:
            title = "======恢复======"
        Timestamp = Sns['Timestamp']
        Message = Sns['Message']
        Message = json.loads(Message)
        try:
            NewStateReason = Message['NewStateReason']
            AlarmDescription = Message['AlarmDescription']
            # 转换cloudwatch单位为友好单位
            datapoint = re.findall(r'[[](.*?)[]]', NewStateReason)
            threshold = re.findall(r'[(](.*?)[)]', NewStateReason)
            count = (datapoint[0].count(","))
            if count == 0:
                datapoint = float(str.split(datapoint[0])[0])
                threshold = float(str.split(threshold[1])[0])
                if threshold > 1000:
                    datapoint = size_b_to_other(datapoint)
                    threshold = size_b_to_other(threshold)
            else:
                i = threshold[len(threshold) - 1]
                pattern = re.compile(r'^[0-9]+.[0-9]')
                if pattern.search(i):
                    threshold = threshold[len(threshold) - 1]
                else:
                    threshold = threshold[len(threshold) - 2] + threshold[len(threshold) - 1]
                
            # 定义消息内容
            content = title + "
    报警主题:" + "" + Subject + "" 
                    "
    报警时间:" + "" + Timestamp + "" 
                    "
    报警原因:" + "" + NewStateReason + "" 
                    "
    友好信息:" + "" + "当前值=" + str(datapoint) + " 连续" + str(count + 1) + "次达到 " + "阀值=" + str(threshold) + "" 
                    "
    备注信息:" + "" + str(AlarmDescription) + ""
        
        except:
            Message = json.dumps(Message, sort_keys=True, indent=2)
            content = title + "
    报警主题:" + "" + Subject + "" 
                "
    详细信息:" + "" + Message + "" 
                "
    备注信息:【消息解析异常】"
    
        data = {
            "msgtype": "text",
            "text": {
                "content": content
            }
        }
        
        data = json.dumps(data)
        request = requests.post(url + token, data, headers=headers)
        return request.text

     测试一下

  • 相关阅读:
    JS鼠标滚动事件
    [Harbor]Harbor简要介绍
    [Kubernetes]yaml文件详解
    [Kubernetes]安装和配置kubectl
    [Docker]如何批量删除镜像
    [Kubernetes]如何使用yaml文件使得可以向外暴露服务
    [Jenkins]CentOS7下Jenkins搭建
    [Docker]Docker拉取,上传镜像到Harbor仓库
    [Harbor]Docker登录Harbor仓库(HTTP方式)
    [Docker]CentOS7通过rpm包安装Docker
  • 原文地址:https://www.cnblogs.com/37yan/p/10709357.html
Copyright © 2020-2023  润新知