Installation Guide for GBase 8s Monitoring Platform
Platform Overview
The unified data platform monitoring and operations system is a database companion tool developed by General Data Technology Co., Ltd. It provides users with visualized monitoring data for database nodes, timely alerting capabilities, intuitive displays of database node status and performance trends, reliable data distribution views, and presentation of database node sessions, system logs, and more.
The unified data platform monitoring and operations system primarily consists of three functional modules: collection agent, collection center, and monitoring website:
Agent: The monitoring data collection agent, deployed on all database nodes from which data is to be collected.
Center: The data collection center, responsible for coordinating data collection across agents, performing logical operations on the collected data, and storing it in the repository.
Website: The monitoring website, providing a monitoring interface for user interaction to accomplish database node monitoring.
Installation and Initialization
Extract the Automatic Installation Package
This platform supports multiple operating systems. Taking Linux as an example, other operating systems use corresponding installation packages. Assume the installation package is GBase8sV8.8Monitor_1.0.0_1_Linux_x86_64.tar, placed in the /opt/gcm_install directory. Extract the compressed package.
Start Service and Connection Configuration
When deploying the web end, a global unified configuration is adopted to enable web deployment and startup, simplifying the process for ease of installation and deployment, and achieving page-based installation and configuration of the collection center and collection agents on the web end, thereby reducing the tedious steps of individual installations.
Step 1: Navigate to the extracted installation package, enter the tomcat bin directory, and execute ./startup.sh to start Tomcat (Tomcat is used in the documentation).
Step 2: After starting the Tomcat service, access the http://[ip]:[port]/gcmonitor page.
Click the "Configure" button to go to the global configuration page.
Step 3: Fill in the relevant parameters in the configuration page as prompted.
Step 4: After filling in, click "Connect" to save the configuration; the backend will save the information and verify the availability of the configuration items.
Step 5: Based on the configured parameters, the background program modifies the properties configuration file of the application.
This completes the initialization configuration of the database.
Enter the unified monitoring website address in the browser, e.g., http://[ip]:[port]/gcmonitor, to reach the user login page:
Enter the username and password (default: admin:admin) and press Enter or click the "Login" button. The system will authenticate the user. If the username and password are correct, you will enter the main unified monitoring interface; otherwise, an error message will appear: Login failed. Please check the username and password!
After successfully logging into the monitoring website, you will enter the main interface:
When initially logging into the monitoring platform, if no collection center and collection agents have been configured, it will prompt that the collection center is unreachable and no data has been collected. If the collection center and agents are all normally connected at this point, the following will be displayed: database node name, last data update time, version information, alarm information, database space utilization, VP busy rate, database node mode, session count, DB Page R/W, and virtual memory usage.
Add GBase 8s Monitoring
Adding GBase 8s monitoring involves three steps: adding an 8s cluster, initializing the collection center, and initializing the collection agent. These steps are completed through the monitoring platform management.
Platform Management
Platform management is primarily used to maintain the basic data for database node monitoring. It includes eight functional modules: Users, Roles, Database Nodes, Collection Center, Monitoring Policies, Alarm Levels, Monitoring Logs, and Monitoring Historical Data Management.
Users can access the platform management interface by clicking "Platform Management" in the main navigation bar:
Create New 8s Cluster
After selecting Platform Management, you can create a new 8s cluster:
Fill in the corresponding information in the pop-up box and choose whether to create a new policy, then finish creating the cluster.
After creating the cluster, you can use the operation bar on the right to initialize the collection center and collection agent.
Initialize Collection Center
Step 1: Go to the Collection Center Management page.
Step 2: Enter the relevant configuration information:
Step 3: After successful saving, automatic installation is performed (the user needs to enter the server connection password before installation).
Step 4: After installation, click the "Start" button to start the center.
Step 5: Check the status in the list to ensure installation and startup are complete.
Initialize Collection Agent
Step 1: First, enter the Collection Agent Service Management page and enter the relevant configuration information:
Step 2: After successful saving, automatic installation is performed (the user needs to enter the server connection password before installation).
Step 3: Click the "Start" buttons for service status and monitoring status to start the collection agent.
With the collection center and collection agent operating normally, after one collection cycle of data is returned, the collection information will be displayed on the cluster status page.
Monitoring Display
Database Overall Status Monitoring
On the cluster supervision page, you can monitor clusters of all instance types within this platform and display the current status of each cluster. Select an individual cluster to view the node status of all its collection centers and collection agents.
Database Node Supervision
Click on the name of an 8s database node in the main database node supervision interface to enter the node summary information page for that database node. The left side displays the 8s database node monitoring navigation bar, as shown below:
This interface comprises three functional parts:
Node Summary Information: The current database status and basic information.
Trend Chart Information: System information trend charts and database information trend charts.
Top Busy Session List Information.
Session Monitoring Summary Information
The session monitoring summary information in session monitoring loads the current active SQL, database transactions, and lock requests into the trend chart in real time.
Top Lock Table
This feature displays the lock status of the top 50 lock tables:
Alarm Information
Alarm information shows different types of alarms and specific error details within the selected time range, including the IP where the error occurred and the corresponding metrics. This page allows you to get an overview of error messages for timely and effective classification and handling based on alarm levels.
Performance Analysis Report
The performance analysis report is generated automatically by the system after selecting the analysis period and the sub-items to be analyzed.
You can view the analysis report for a specific database, specified time period, and selected sub-items. Click "Result Analysis" to display the specific analysis report and the corresponding metric trend charts below.
above, after selecting the CPU usage trend chart, the analysis reveals the table with the highest database activity, along with the corresponding average CPU busy rate, database busy rate, and operating system average load trend chart, providing clearer insight into the situation during a specific past period.
This platform also supports exporting the selected analysis report as Word and PDF documents.
Summary
The unified data platform monitoring and operations system supports monitoring of single or multiple database nodes. By collecting information such as system resource utilization, network communication status, process running status, and database node operation status, and calculating based on user-defined alarm strategies, it notifies users of alarms in a timely manner to prevent database node failures. It supports persisting collected information to the database and provides performance analysis and display from multiple dimensions, greatly facilitating database node tuning and troubleshooting.