As the saying goes, "When the human heart is not enough, the snake swallows the elephant." After meeting the basic requirements in the previous article, more advanced functional requirements need to be met.
1. Simultaneous monitoring of multiple locations and lines
Multiple monitoring locations refer to monitoring websites and servers in different places at the same time, such as monitoring at multiple points such as Guangzhou, Beijing, Chongqing, etc., and an alarm will be issued if a fault is detected at any point. Multi-line refers to using different lines (Telecom, China Unicom, China Railway Communications, etc.) for simultaneous monitoring. Location and line are used together, such as Guangzhou Telecom, Beijing Unicom, etc.
China's Internet is artificially divided into north and south. The south is mainly Telecom, and the north is mainly China Unicom (formerly China Netcom), as well as other education networks, China Railcom, etc. The access speed of different network lines is slower and more unstable than that of the same network line. For example, China Telecom's access to China Unicom is several times slower than China Telecom's access to China Telecom, and the interconnection may not work from time to time. Coupled with the different locations, the situation is even more complicated. For example, pinging Guangzhou Telecom in Guangzhou usually takes 10ms, pinging Guangzhou China Unicom usually takes 30ms, pinging Beijing Telecom takes 20ms, and pinging Beijing Netcom takes 60ms. In Guangzhou, we often monitor that Guangzhou Telecom lines cannot access Guangzhou Unicom lines.
Most webmasters and network administrators do not have the conditions to monitor multiple locations and lines themselves. Online monitoring can easily help you solve this problem. If your website server is nationwide, it requires multiple locations and line monitoring.
2. Fault alarm supports major alarm and minor alarm
Servers generally open multiple services at the same time, such as website services, email services, FTP services, database services, etc. Network administrators like to be able to monitor whether these services are normal at the same time. If there is no prioritization, when the server has a network failure or is restarted for maintenance, you will immediately receive website failure alarms, email service alarms, and FTP service alarms. Even worse, when the server cannot be restored for a short period of time, alarm messages will continue to flow. Sent to you until the server comes back to normal or you crash.
For services that monitor the same server, a primary monitoring object needs to be defined. When the primary monitoring object fails, only the alarms of the primary monitoring object are sent, and other secondary monitoring objects suspend monitoring and alarming. For example, Ping is used as the main monitoring object. If Ping fails and Timeout occurs, it means that the server has crashed or is disconnected. At this time, only server Ping alarms are sent to continuously monitor Ping, because it is no longer necessary to continue monitoring and alarming other services. This can greatly reduce the number of alarm messages and make monitoring more reasonable and efficient.
3. Customize alarm conditions
The general monitoring result is success or failure, such as Ping failure, webpage access error, and Socket failure. When these occur, they are called faults, and faults are the highest priority alarms. In addition, the return delay and content can also be monitored, such as the Ping return delay, the time of accessing the web page, the content obtained by accessing the web page, etc. The returned results can be used to customize alarm conditions. For example, the return delay of Ping monitoring is generally between 10-30ms. When the delay is greater than 100ms, it means that there may be a problem with the network or server, causing slow network response, and it is necessary to immediately check whether there is traffic. Issues such as being too large or the server CPU being too high.
Customized alarm conditions can further refine the monitoring content, accurately know the status of the website and server, and gain more time to solve the problem before the failure occurs.
4. Customize alarm message content
There are many alarm message contents, such as alarm website name, server IP, monitored line, monitored service, error message, occurrence time, etc. You can customize the alarm message content to freely select the received alarm content and modify it to suit your actual situation. Require. This is particularly meaningful for using SMS to receive alarm content. The SMS content can be up to 70 characters. It is difficult to fully understand the fault content within 70 characters. It is also necessary to customize the content and adjust it so that the fault can be known even with limited character content. Time, fault content, such as: "007电影网12:32 Telecommunications line monitoring failed for the first time", clearly know the fault information.
5. Receive summary reports via email
When there are no major problems with the website and server, no one is too lazy to log in to the online monitoring service management backend to check. If you can receive a summary report email of website server monitoring every day, you can get a general understanding of the website and server status in two or three minutes.
6. Passive monitoring
Although active monitoring does not require the installation of codes and programs and is very safe and convenient, it lacks a lot of detailed monitoring content, such as the inability to obtain hard disk size, CPU usage, network traffic, etc. These monitoring content are very useful, such as the CPU is too high It means there is a problem with the website or program, and high traffic means it may be attacked, etc.
SNMP (Simple Network Management Protocol) is commonly used for passive monitoring. Most of the content you are interested in can be monitored through SNMP. Most operating systems support SNMP, which makes activation and management very convenient and very secure. The disadvantage of SNMP is that it takes up bandwidth and consumes a certain amount of CPU and memory. When the CPU is too high and the network traffic is heavy, it cannot be effectively monitored.
[Articles related to the author]
Discussing the necessity of using online website server monitoring (1)
Discussing the necessity of using online website server monitoring (2)
The author of this article: Zhongyu Internet Applications and Services is original. Please leave the source of the article when reprinting it.
Thanks to Zhongyu Internet for your contribution