How to monitor instances through ECS's built-in monitoring service and cloud monitoring service

The content of this article is about how to monitor instances through the ECS built-in monitoring service and cloud monitoring service. It has certain reference value. Friends in need can refer to it. I hope it will be helpful to you.

Monitoring

It is very important to monitor the health of your ECS instances. You need to ensure that users can always quickly open your website and applications, or quickly complete tasks such as data processing and rendering. Alibaba Cloud provides services such as monitoring data collection, visualization, and real-time monitoring and alarming to ensure that your instances are always in normal running status.

Monitoring details

Currently, you can monitor instances through the ECS built-in monitoring service and cloud monitoring service. ECS's own monitoring service provides vCPU usage, network traffic, and disk I/O monitoring. Cloud monitoring provides more refined monitoring granularity. For more details, please refer to the monitoring item description. The following is an analysis of some monitoring information:

vCPU: Alibaba Cloud provides instance vCPU usage monitoring data, the unit is percentage. The higher the percentage value, the higher the instance vCPU load. You can query monitoring data through the ECS management console, cloud monitoring management console, calling the ECS API, or remotely connecting to the instance. The following is how to view vCPU usage after connecting to an instance remotely:

Windows instance: View vCPU usage in Task Manager. You can sort by vCPU usage to locate the processes occupying the instance vCPU resources.

Linux instance: Run the top command to view vCPU usage. Press Shift P on the keyboard to sort by vCPU usage and locate the processes occupying the instance vCPU resources.

Network traffic: Alibaba Cloud provides network traffic monitoring data in the outbound and inbound directions of the instance, in kbps. ECS's own monitoring service generally provides public network traffic monitoring, and cloud monitoring can obtain public network and intranet traffic monitoring. For example, your public network outbound bandwidth is 1 Mbps. When the outbound traffic reaches 1024 kbps, it means that your public network bandwidth is fully loaded.

ECS comes with monitoring service

The steps to view monitoring information on the ECS management console are as follows:

Log in to ECS Management console.

In the left navigation bar, click Instances.

Select a region.

Find the target instance and click the instance name.

On the instance details page, you can see monitoring information, including vCPU usage and inbound and outbound network traffic.

Click the icon to set the monitoring time period.

Description

Due to the different display aggregation methods, the length of the selected time period will affect the accuracy of the display. The smaller the selected time range, the more detailed the display effect will be. For example, 5-minute and 15-minute averages will show different results.

(Optional) Click Set Alarm Rules to go to the cloud monitoring console to set vCPU usage and network traffic alarm rules. See the Alarm Services Overview for more details.

(Optional) Click to view more indicators such as memory to go to the cloud monitoring console to collect more data. After a moment, the monitoring data will be updated to the latest status.

You can also use the ECS API DescribeInstanceMonitorData, DescribeDiskMonitorData, and DescribeEniMonitorData to obtain monitoring data.

The following is a detailed list of monitoring items provided by ECS. The indicator collection granularity is 1 minute.

Cloud Monitoring

Cloud Monitoring provides you with an enterprise-level open one that can be used out of the box. Stationary monitoring solution. Cloud Monitor provides host monitoring services for your ECS. For more details, please see Host Monitoring Overview. The following steps demonstrate how to obtain ECS instance monitoring data from the Cloud Monitoring Management Console.

Log in to the cloud monitoring management console.

In the left navigation bar, click Host Monitoring.

Find the target instance.

(Optional) If the cloud monitoring plug-in is not installed on the instance, click Install to update the plug-in.

Click on the monitoring chart to obtain monitoring data.

Click Alarm Rules to set alarm rules.

Appendix: Knowledge related to bandwidth units

The difference between Kb and KB

Information in computers is represented by binary 0s and 1s. Each 0 or 1 is called a bit, represented by a lowercase b. 8 bits constitute 1 byte (Byte), similar to 0101 0010, represented by a capital B, 1 Byte=8 bits (1B=8b).

Use uppercase K or lowercase k to indicate thousands of hours, Kb is thousands of digits, and KB is kilobytes.

In the network traffic monitoring that comes with ECS, ps refers to /s, that is, per second. kbps refers to network speed, that is, how many kilobits of information are transmitted per second. Typically, bps is omitted when describing bandwidth. For example, the complete writing of 4M bandwidth should be 4 Mbps.

The relationship between bandwidth and download speed

Myth: The download speed is the same as the bandwidth.

Correct: Taking 1Mbps bandwidth as an example, 1KB=8Kb, 1Mbps=125KB/s, 1kbps=1000bps

The download rate of 1Mbps bandwidth is theoretically 125KB/s, but in some examples Applications will occupy a small amount of bandwidth, such as remote connections, so the actual download rate is often 100 KB/s~110KB/s.

The above is the detailed content of How to monitor instances through ECS's built-in monitoring service and cloud monitoring service. For more information, please follow other related articles on the PHP Chinese website!