客户端需求
客户端说白了就是存储一些插件脚本。然后通过服务端传来的配置文件（监控项、监控项对应的监控间隔）、
在本地定时的去执行每个监控项对应的监控插件。
并且汇报数据。
客户端目录
MonitorClient.py
启动main
setting.py
配置文件。记录主机的ID号，服务端的地址、端口、API的url、发送数据超时时间、配置文件更新时间。
configs ={ 'HostID': 1, "Server": "localhost", "ServerPort": 9000, "urls":{ 'get_configs' :['api/client/config','get'], #acquire all the services will be monitored 'service_report': ['api/client/service/report/','post'], }, 'RequestTimeout':30, 'ConfigUpdateInterval': 300, #5 mins as default }
main.py
启动脚本。
import client class command_handler(object): def __init__(self, sys_args): self.sys_args = sys_args if len(self.sys_args)<2:exit(self.help_msg()) self.command_allowcator() def command_allowcator(self): '''分捡用户输入的不同指令''' print(self.sys_args[1]) if hasattr(self,self.sys_args[1]): func= getattr(self,self.sys_args[1]) return func() else: print("command does not exist!") self.help_msg() def help_msg(self): valid_commands = ''' start start monitor client stop stop monitor client ''' exit(valid_commands) def start(self): print("going to start the monitor client") #exit_flag = False Client = client.ClientHandle() Client.forever_run() def stop(self): print("stopping the monitor client")
client.py
客户端逻辑处理
class ClientHandle(object): def __init__(self): self.monitored_services = {} def load_latest_configs(self): ''' load the latest monitor configs from monitor server :return: ''' request_type = settings.configs['urls']['get_configs'][1] url = "%s/%s" %(settings.configs['urls']['get_configs'][0], settings.configs['HostID']) latest_configs = self.url_request(request_type,url) latest_configs = json.loads(latest_configs) self.monitored_services.update(latest_configs) def forever_run(self): ''' start the client program forever :return: ''' exit_flag = False config_last_update_time = 0 while not exit_flag: if time.time() - config_last_update_time > settings.configs['ConfigUpdateInterval']: self.load_latest_configs() print("Loaded latest config:", self.monitored_services) config_last_update_time = time.time() #start to monitor services for service_name,val in self.monitored_services['services'].items(): if len(val) == 2:# means it's the first time to monitor self.monitored_services['services'][service_name].append(0) monitor_interval = val[1] last_invoke_time = val[2] if time.time() - last_invoke_time > monitor_interval: #needs to run the plugin print(last_invoke_time,time.time()) self.monitored_services['services'][service_name][2]= time.time() #start a new thread to call each monitor plugin t = threading.Thread(target=self.invoke_plugin,args=(service_name,val)) t.start() print("Going to monitor [%s]" % service_name) else: print("Going to monitor [%s] in [%s] secs" % (service_name, monitor_interval - (time.time()-last_invoke_time))) time.sleep(1) def invoke_plugin(self,service_name,val): ''' invoke the monitor plugin here, and send the data to monitor server after plugin returned status data each time :param val: [pulgin_name,monitor_interval,last_run_time] :return: ''' plugin_name = val[0] if hasattr(plugin_api,plugin_name): func = getattr(plugin_api,plugin_name) plugin_callback = func() #print("--monitor result:",plugin_callback) report_data = { 'client_id':settings.configs['HostID'], 'service_name':service_name, 'data':json.dumps(plugin_callback) } request_action = settings.configs['urls']['service_report'][1] request_url = settings.configs['urls']['service_report'][0] #report_data = json.dumps(report_data) print('---report data:',report_data) self.url_request(request_action,request_url,params=report_data) else: print("\033[31;1mCannot find service [%s]'s plugin name [%s] in plugin_api\033[0m"% (service_name,plugin_name )) print('--plugin:',val) def url_request(self,action,url,**extra_data): ''' cope with monitor server by url :param action: "get" or "post" :param url: witch url you want to request from the monitor server :param extra_data: extra parameters needed to be submited :return: ''' abs_url = "http://%s:%s/%s" % (settings.configs['Server'], settings.configs["ServerPort"], url) if action in ('get','GET'): print(abs_url,extra_data) try: req = urllib2.Request(abs_url) req_data = urllib2.urlopen(req,timeout=settings.configs['RequestTimeout']) callback = req_data.read() #print "-->server response:",callback return callback except urllib2.URLError as e: exit("\033[31;1m%s\033[0m"%e) elif action in ('post','POST'): #print(abs_url,extra_data['params']) try: data_encode = urllib.urlencode(extra_data['params']) req = urllib2.Request(url=abs_url,data=data_encode) res_data = urllib2.urlopen(req,timeout=settings.configs['RequestTimeout']) callback = res_data.read() callback = json.loads(callback) print "\033[31;1m[%s]:[%s]\033[0m response:\n%s" %(action,abs_url,callback) return callback except Exception as e: print('---exec',e) exit("\033[31;1m%s\033[0m"%e)
load_latest_configs这个函数是获取服务端的配置文件
forever_run是以多线程的方式定期执行插件脚本
url_request定义了具体的提交方式
invoke_plugin是从插件脚本中获取数据，添加到字典中调用url_request发送给服务端
plugins
这个目录是插件目录
windows和linux子目录存有对应系统的插件脚本
plugin_api.py存着监控项（接受自服务端）和插件脚本的对应关系。