Common experience
Talk about some of the work often encountered some time, perhaps you have met, happy to work, just opened the computer, the following situation:
The leader ran up to ask yesterday website visit very slow, server again problem
Customer service run for xx city in Fujian province, users said that the website opened very slowly, the server has a problem
The boss said he couldn't open the website at home yesterday and the server went wrong.
The technical director said he just got on the CDN yesterday and you see how it works.
The sales department asked if we could see how fast we visited our website and how to improve
There are more about the website operation and maintenance of the blame, welcome to enumerate ...
Why is the problem always considered to be the cause of operation and maintenance?
Say a digression, in a company unexpectedly met former colleague, meet greeting a few words, he said a sentence let me so far unforgettable words: "Or you yun-dimensional relaxed, every day what is not to do, as long as staring at the screen is good, staring at the server is not a problem."
Reasons for slow website access
Server failure
Problem with program logic, resulting in slow response
The page is slow for an element, causing the whole page to slow
Slow User network environment
South Power North Network slow
The trouble of operation and maintenance
Some people propose to use Zabbix to try, as a single node operation and maintenance Monitoring tool, Zabbix really powerful, but it does not have the full stack of network performance monitoring, you think Zabbix is the big God, hehe, I can not know the past things, such as the situation of the visit I will not know, I have only one server. Some people say use the Web test software to try it, but he is just a normal get, but the egg.
Workaround
Eventually these impact site access problems can be well resolved, cloud wisdom of the monitoring treasure is a good choice, the page performance management and site monitoring can let you get rid of unnecessary blame. Words do not say much, on several figures:
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100236zsrkrkroksztkjzd.png "/>
Comparison of monitoring points
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100250dtj57ti4qykfuy7v.png "/>
Web page Performance Management
China's dozens of provinces, Wuhan Telecom bottom.
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100302hy8333j337f46frj.png "/>
Page open time data for all monitoring points is currently listed, and you can see performance scores and response times for each region
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100426mpk846kq33p6auqe.png "/>
Timing Diagram
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100442sam2owcfw1vn5f51.png "/>
Response time for each resource
The Web page is slow, it is possible that some elements of the Web page down, you can monitor the page loading of various elements (used Firebug know), we can know the DNS resolution time, establish a connection, send requests, wait, receive data consumed time, and Firebug basically exactly the same. We can see that the various resources of the time are listed in detail, we can very accurately analyze the problem in which network link.
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100527ghni4uehsuthmn8u.png "/>
Web page Performance Management-Request/Response headers
You can see the server response header, the general head contains file expiration time, cache hit situation, and so on, are some of the information to help troubleshoot problems.
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100545tmpwywi41yfpjvw2.png "/>
Site Availability Rate
Get the availability of a day, you can see the Shanghai Science and Education network availability rate of 75%, have not heard, the low availability rate may be a matter of course.
How it's achieved.
Monitoring Bao provides Web performance management This function, only need simple configuration. Login to the background, click "Monitoring"-"Web performance management-" to create a monitoring project.
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100621iiojcb7lpc4ynkqn.png "/>
Create a monitoring item
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100659dceuleut6vwccf5d.png "/>
Detecting nodes
Cloud Wisdom presented to the operation of the survival time account, a total of more than 30 monitoring nodes to choose from, the Enterprise Edition account can be selected throughout the country and overseas major cities more than 100 monitoring points, including various regions, various networks. Monitoring frequency selection 15 minutes, the smaller the frequency of the more abundant data.
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100730dk3eppugezu0zz7c.png "/>
Alarm Configuration
OPS can customize the alarm trigger according to the SLA of its own business, for example, if any one node response time exceeds 5000ms to send an alarm, alarm mode has email, SMS, and telephone voice. You can choose the appropriate alarm mode based on the severity of the alarm condition.
650) this.width=650; "Src=" https://dn-linuxcn.qbox.me/data/attachment/album/201509/23/100749xv4m4v6nab554zfb.png "/>
Detection configuration Complete
At last
If you are being plagued by a variety of web site operation and maintenance problems, try to monitor the performance of the Web page, not only to meet the boss, leaders, colleagues of all kinds of pit daddy needs, but also the first time to discover the server and network failure, the user complaints to eliminate in the embryonic state, henceforth no longer blame.
Operation and maintenance Survival time hematemesis: How to get rid of Web site operation and maintenance