Objective:
General server will be built monitoring, since the monitoring has, then the alarm is definitely necessary. Monitoring whether it is CACTI,NAGIOS,ZABBX, etc., alarm regardless of fetion,, mail, SMS. As long as the appropriate administrator, to help them grasp the status of the server in time, it will certainly save a lot of things. The author uses the mail monitoring, but also is offsite. And then set up a proxy alarm in QQ. That is, mail can be received, as long as QQ or all online, you can also receive, it is convenient to use.
Problem:
Things are not as good as you think!!! Recently, always intermittent email, especially depressed. Today, however, it took a morning to troubleshoot. It was a ridiculous question that led to it. Network bandwidth!
Alarm: (Simple cut part of the figure)
650) this.width=650; "Src=" Http://s2.51cto.com/wyfs02/M00/82/76/wKioL1dWPzei_pVYAAA8Hry_Hd4068.jpg-wh_500x0-wm_3 -wmp_4-s_1520648379.jpg "title=" monitoring. jpg "alt=" wkiol1dwpzei_pvyaaa8hry_hd4068.jpg-wh_50 "/>
Troubleshooting: (When you want to check the server first)
1, check whether the service of the two sides is normal operation. (Netstat,ps, etc.).
2, check the monitoring inside the configuration is correct (IP, password and other configuration).
3. Check whether the server iptables firewall and corporate firewall have policy blocking.
4, check the log, whether the monitoring log normal output.
5, check the server hardware firewall is normal (although intermittent, but after all, it can be monitored, estimating the relationship is not small).
。。。。。。
However, there is nothing abnormal!
Suddenly remembered, remember some time ago always someone said the net is not good what, is it related?
Immediately landing the router see configured those things, the company is a 100Mbps link. To DHCP plus some specific configuration server, there are still 50Mbps, but many of the company's people will configure their own TCP/IP parameters, so many people's IP is not within the DHCP, then there is no speed limit for these people, of course, my server's IP is not within the DHCP, Did you think it was the problem?
Continue troubleshooting:
1, local server first monitor with local traffic. (such as iftop, self-monitoring, etc., the traffic does not go much)
2, practical curl to obtain the company's website, whether it can be obtained (results are not available, a page needs about 1000kbps)
Curl: (7) couldn ' t connect to host
So, why not change the bandwidth?
Then, the bandwidth of the monitoring was changed to 3000kbps
Then the utility curl, can get the page normally, within 10 minutes, the monitoring mail also normal!
Well, for the time being the net problem! That's hilarious! Write a blog and tease.
This article comes from "? Only! "Blog, be sure to keep this provenance http://renzhiyuan.blog.51cto.com/10433137/1786905
Monitoring, why do you always send emails on and off?