Viewing Performance Metrics of a GaussDB(DWS) Cluster

The Performance Monitoring page displays the resource consumption trends of clusters, databases, nodes, instances, and logical clusters. After selecting a time range, you can view the performance trend in the time range. GaussDB(DWS) allows you to view the monitoring data of the last seven days. By default, monitoring data of the last hour is displayed. Performance monitoring also allows data aggregation of different periods. You can aggregate raw data based on the corresponding sampling period to display indicator trends of a longer period.

Performance Monitoring

Log in to the DWS console.
Choose Dedicated Clusters > Clusters and locate the cluster to be monitored.
In the Operation column of the target cluster, click Monitoring Panel.
In the navigation pane on the left, choose Monitoring > Performance Monitoring.

Notes and Constraints

A maximum of 20 views can be added to each panel. Adding too many views will increase the number of page requests and the rendering time.
A maximum of 20 monitored objects can be selected in the node dimension. This feature is supported only in 8.1.3.310 and later cluster versions.

Monitoring Panel Operations

You can configure monitoring views by customizing monitoring panels. Monitoring panels are bound to users. After logging in to the system, you can view the user-defined monitoring panels. Table 1 lists the operations.

Figure 1 Monitoring Panel
Click to enlarge

**Table 1** Monitoring Panel operations
Operation	Description
Create Monitoring Panel	You can click Create Monitoring Panel to customize a monitoring panel.
Modify	You can click Modify to change the name of a monitoring panel.
Share	You can click Share to share a monitoring panel. The recipients can view the panel but cannot modify it.
Delete	You can click Delete to delete a monitoring panel. The default monitoring panel cannot be deleted.

Adding a Monitoring View

Currently, DMS provides monitoring views for clusters, databases, instances, nodes, and logical clusters. You can click Add View to add a monitoring view as required. Table 1 Monitoring view metrics lists the monitoring metrics.

Figure 2 Adding a Monitoring View
Click to enlarge

**Table 2** Monitoring view metrics
Monitoring View Dimension	Metric Name	Description	Monitoring Interval (Raw Data)
Cluster	CPU Usage	Ratio of the time taken by tasks on the CPU to the total available time in a cluster	60s
	Memory Usage	Ratio of the allocated and used memory to the total memory in a cluster	60s
	Disk Usage	Ratio of the used disk space to the total disk space	30s
	Disk I/O	Total amount of data read from or written to a disk per second in a cluster	60s
	Network I/O	Amount of data transmitted over the network per second in a cluster	60s
	Status	Cluster status	60s
	Abnormal CNs	Number of abnormal CNs	600s
	Read-only	Cluster read-only status	60s
	Sessions	Number of sessions in a cluster	180s
	Active Sessions	Number of queries in the active or fastpathfunctioncall state in a cluster	180s
	Deadlocks	Number of deadlocks in a cluster	180s
	Abnormal DNs	Number of abnormal DNs	600s
	DN CPU Usage	CPU usage of the DN instance	60s
	Transactions per Second	Average number of transactions processed per second	60s
	Queries Per Second	Average number of user queries per second	60s
	Capacity	Cluster capacity	86400s
	Schema Capacity	Capacity of the cluster schema	3600s
	Stacked SQL	Number of SQL statements in the waiting state in a cluster	60s
	Queries	Number of user queries	60s
	Resource Pool	Resource pool used by the user	120s
	Transactions	Number of transactions processed by a user	60s
Database	Active Sessions	Number of queries in the active or fastpathfunctioncall state in a database	180s
	Sessions	Number of database sessions	180s
	Number of Inserted Rows	Number of rows inserted by queries in this database	60s
	Number of Updated Rows	Number of rows updated by queries in this database	180s
	Number of Deleted Rows	Number of rows deleted by queries in this database	180s
	Capacity	Database capacity, in bytes.	86400s
Instance	Instance Memory Usage	Memory usage of an instance	60s
Instance	Instance Dynamic Memory Usage	Dynamic memory usage of an instance	60s
Node	CPU Usage	Ratio of the time taken by tasks on the CPU to the total available time of a node	60s
	Memory Usage	Ratio of the allocated and used memory to the total memory of a node	60s
	Average Disk Usage	Ratio of the disk read/write time to the total time in a unit time	30s
	TCP Protocol Stack Retransmission Rate	Ratio of the number of times that TCP data packets on a node need to be resent to the total number of times that TCP data packets are sent within a specified period	60s
	Disk I/O	Total amount of data read from or written to a disk of a node per second	60s
	Network I/O	Amount of data transmitted over the network per second on a node	60s
	Total Disk Space	Total disk space of a node	30s
	Disk Usage	Ratio of time when the disk processes read and write requests in a specified period	30s
	Disk Read Rate	Rate at which data is read from the device per second	30s
	Disk Write Rate	Rate at which data is written to the device per second	30s
	Disk I/O Wait Time	Average waiting time for each I/O request	60s
	Disk I/O Service Time	Average processing time for each I/O request	60s
	Disk I/O Usage	Ratio of time when the disk I/O processes read and write requests in a specified period	60s
	NIC Status	NIC status (up/down)	60s
	Received Packets	Number of packets received by the NIC	60s
	Sent Packets	Number of packets sent by the NIC	60s
	Lost Packets Received	Number of received data packets that are dropped by the NIC	60s
	Receive Rate	Rate at which data is received by the NIC	60s
	Transmit Rate	Rate at which data is sent by the NIC	60s
	CPU Usage	CPU usage	60s
	Memory Usage	Memory usage	60s
Logical cluster	CPU Usage	Ratio of the time taken by tasks on the CPU to the total available time in a logical cluster	60s
	Memory Usage	Ratio of the allocated and used memory to the total memory in a logical cluster	60s
	Disk I/O	Total amount of data read from or written to a disk per second in a logical cluster	60s
	Network I/O	Amount of data transmitted over the network per second in a logical cluster	60s
	QPS	Number of requests that can be processed in a unit time	60s
	Stacked SQL	Number of SQL statements in the waiting state in a logical cluster	60s

Exporting Monitoring Data

Performance Monitoring supports data export. You can click Export Data to further process data. By default, data in all monitoring views on the current page is exported. The export time range is subject to the selected time range.

Click to enlarge