Overview-Scalable Open Financial Architecture(SOFAStack)-阿里云帮助中心

The application monitoring feature of Real-time Monitoring Service (RMS) allows you to analyze applications. You can view and analyze all application monitoring data on the Application Monitoring page. After an application is deployed by using Cloud Application Fabric Engine (CAFE), RMS automatically monitors the application and displays monitoring data of the application on a dashboard in various dimensions from various perspectives. This helps meet the monitoring needs of the R&D or O&M engineers of the application to ensure the stability of the application.

Background information

The application monitoring feature of RMS can monitor container applications and classic applications. Application metrics are divided into the following four dimensions:

Dimension	Example	Description
Service metrics	Error, Service, SAL, CAL, and DAL	Framework configuration: For applications developed based on the Scalable Open Financial Architecture (SOFA) or SOFABoot framework, metrics such as PV, Service, SAL, SQL, CAL, and CE Thread are automatically collected based on the log file paths in the default framework. You need only to enable monitoring for the applications. For more information about the log file paths in a specific framework, see Service metrics. Custom monitoring configuration: For applications that are not developed based on the SOFA or SOFABoot framework, you must configure a log file path and column value for each metric, such as Error or DAL. For applications developed based on the SOFA or SOFABoot framework, if you do not want to use the default log file path, you can switch the configuration mode to custom monitoring configuration.
System metrics	CPU, memory, and disks	The system metrics of container applications are collected by Apache System Activity Reporter (ASAR) from containers on a server, whereas the system metrics of classic applications are collected by ASAR from processes on a server.
Java virtual machine (JVM) metrics	garbage collection (GC) and threads	After the process of a Java application is started, a file named after the ID of the process is created in the `/tmp/hsperfdata_${username}` directory, and JVM information is stored in the file. The ${username} variable specifies the current username, and the information about all started Java processes of the user is stored in the `/tmp/hsperfdata_${username}` directory. The JVM metrics of RMS are collected by parsing the `/tmp/hsperfdata_${username}/${pid}` file, such as `/tmp/hsperfdata_admin/1234`.
Port detection	/	Port liveness detection is port detection initiated from the agent for an application service. For example, a service exposes port 12200 and can be pinged under normal conditions. However, if the service is down, the port cannot be accessed. Port detection can periodically check whether an application service is alive to ensure that the application can provide external services.
Port detection	/	Port liveness detection is to initiate port detection for the application service from the agent. For example, a service exposes port 12200. Under normal circumstances, it can be connected by running the ping command, but if the service is in downtime, the port cannot be connected. The survival status of the application service can be periodically determined through port detection to ensure that the application can provide external services.

To ensure the high availability of applications, you can deploy multiple instances or logical data centers (LDCs). Different application instances, namely, single machines, may be distributed across different Internet data centers (IDCs) and different LDCs. Therefore, RMS aggregates metric data in each dimension from the single machine, IDC, and LDC perspectives, and performs comparative analysis on the data based on spatial distribution and temporal distribution of IDCs, LDCs, and single machines.

Procedure

Log on to the RMS console.
In the left-side navigation pane, click Application Monitoring.
The Application Monitoring page displays the summarized information of applications on the My Applications and All Applications tabs.
- All Applications: After you create and publish applications by using CAFE, RMS synchronizes all published applications to the All Applications tab.
- My Applications: On the All Applications tab, find an application and choose More > Add to Favorites to add the application to the My Applications tab.

All Applications

Click the All Applications tab to view the following information:

Category	Description
Overview information	Total of applications Number of alert applications
List information	Application name Metric alert status Number of alert rules Alert metric not configured Alarm switch Actions: Edit: Click Edit to jump to the application configuration interface. More: Collecting applications, subscribing to alerts, viewing alert history, viewing alert rules, and pausing alerts.

My applications

Click the My Applications tab to view the following information:

Category	Description
Overview information	Total of followed applications Number of healthy applications Number of exceptional applications Health: The proportion of healthy applications to total applications.
parameters	Application name Metric alert status Number of alert rules Alert metric not configured Alarm switch Actions: subscribing to alerts, viewing alert history, and viewing alert rules.

You can also perform the following operations on this page:

Batch adding followed applications: You can batch add followed applications.
Batch canceling followed applications: You can batch cancel some followed applications.
Duplicating followed applications: You can duplicate followed applications for subsequent processing, e.g., send them to friends.