Server alert mechanism configuration tips

Today's servers are equipped with piles of sensors. Once any sensor happens, an alarm is triggered, Which is dizzy. However, administrators must be aware that the alarms generated by server monitoring tools are not always important. If the server sends an alarm at will, you will be forced to receive a large amount of information within a short period of time, and the truly important information may be drowned. The purpose of this article is to help administrators identify which server alarms are important. administrators need server monitoring tools to correctly provide alarms.

Tips for configuring alert mechanisms

Before starting, I would like to point out that there is no absolutely correct or incorrect method for configuring alarms. The suggestions in this article are mainly based on my 20 years of IT experience, but they are my personal preferences. Therefore, although I hope you can get the benefits from my suggestions, I suggest that every administrator configure Server alarms to keep up with the specific requirements of his company as close as possible.

Another point I should point out is that the administrator can use many methods to generate server alarms. Some servers can generate alarms at the hardware level. Although this is very useful, it is far from meeting the requirements of a dedicated alarm mechanism. The server monitoring tool provided by the server supplier has a large amount of information. It is used as an operating System-level server monitoring tool, such as Microsoft's System Center Operations Manager. Because there are many choices for monitoring tools, I will generally talk about them, rather than talking about a specific tool separately.

Configure priority for the server alert mechanism

To make server monitoring more effective, the key is to configure the priority of the alarm mechanism of the monitoring tool. I recommend that you divide alarms into three types: advanced, intermediate, and low.

I usually define high-priority alerts as the most critical transactions. For example, exhausting all the disk space on the server is a critical event. Failure of the same cluster application server can cause a serious disaster.

It is difficult to define alerts with intermediate priority. In my opinion, intermediate warnings may be advanced for other organizations. For me, if problems in the alarm do not cause overall downtime, the priority is intermediate. For example, if a node in a cluster does not know why the power is down but does not affect the operation of the entire cluster, I will define it as a priority intermediate. Of course, this is closely related to the working environment. Some big companies I have worked on define such problems as critical transactions.

However, if your company cannot tolerate any downtime, the clever way is to determine the priority based on whether the problem has the potential risk of failure. For example, if you have a RAID array that can handle errors of Two Disks without going offline, you can use this as a medium-priority alarm if one of them fails, because the array can still handle the fault of another disk. However, if both disks fail, it is a higher priority, because if there is another disk failure, the entire array will be complete.

Although I think this method is easy to prioritize, it is more difficult to configure alerts based on the number of parts than to trigger alerts in case of a fault. Depending on the type of monitoring you use and the available functions of the monitoring software, it is worth choosing to install such an alarm system.

Configure Monitoring Mechanism

Once you decide how to classify alerts, you need to decide how to notify alerts. I personally prefer to ask the server monitoring tool to send high-priority alarms to me via mobile phone information. It is the best way to send key alarms to my mobile phone as soon as possible because I carry them with me on my mobile phone.

Although medium-priority alerts are important, they are not absolutely critical, so I tend to send them by email ., Windows Server has the email sending function, which means that you can easily send email alerts based on the time when the operating system occurs.

Windows can automatically send email alerts

I check emails every day, so alarms will not be ignored. I may not want to know the alarm through my mobile phone. It is very important to differentiate between middle and high levels. I don't want to disturb my friends when they celebrate the weekend. Of course, this is only an example of alarm, and there are other options. For example, a company named Server Density installed an iPhone Server monitoring application that supports all alarms.

Of course, the object of higher priority remains to be discussed. Another thing to consider is that high priority alarms are not necessarily related to system faults. Most servers trigger alarms as long as the system case is enabled. If only you can open the server case, the case alert is obviously a higher priority alert. Likewise, alarms with high temperatures are of high priority, because server overheating will eventually lead to shutdown.