Viewing the Instance Monitoring Page

Scenarios

On the HetuEngine web UI, you can view the detailed information about a specified service, including the execution status of each SQL statement. If the current cluster uses dual planes, a Windows host that can connect to the cluster service plane is required.

You cannot view the compute instance task monitoring page using Internet Explorer.

Prerequisites

You have created an administrator for accessing the HetuEngine web UI. For details, see Creating a HetuEngine User.

Procedure

1. Log in to FusionInsight Manager as an administrator who can access the HetuEngine web UI and choose Cluster > Name of the desired cluster > Services > HetuEngine. The HetuEngine service page is displayed.
2. In the Basic Information area on the Dashboard tab page, click the link next to HSConsole WebUI. The HSConsole page is displayed.
3. Locate the row of the target Instance ID, and click LINK in the WebUI column of the row. The compute instance task monitoring page is displayed on a new page. The Query History page is displayed for the first access. Click Metrics to view the compute instance task monitoring page.
   Figure 1 Compute instance task monitoring page
   

   Table 1 Metric description

   Metric	Description
   Cluster CPU Usage	Indicates the CPU usage of the current instance.
   Cluster Free Memory	Indicates the free memory of the current instance.
   Average Cluster CPU Usage	Indicates the average CPU usage of the current instance.
   Used Query Memory	Indicates the memory used by the current instance.
   Running Queries	Indicates the number of tasks concurrently executed on the current instance.
   Queued Queries	Indicates the number of tasks to be executed in the waiting queue on the current instance.
   Blocked Queries	Indicates the number of blocked tasks on the current instance.
   Active Workers	Indicates the number of valid Workers on the current instance.
   Avg Running Tasks	Indicates the average number of running tasks in the current instance.
   Avg CPU cycles per worker	Indicates the average CPU cycles of each Worker on the current instance.

4. Filter query tasks by State on the page.
   Figure 2 Filtering tasks by State
   

   Table 2 State description

   State	Description
   Select All	Views tasks of all statuses.
   Queued	Views the tasks to be executed in the waiting queue.
   Waiting For Resources	Views the tasks that wait for resources.
   Dispatching	Views the tasks that are being dispatched.
   Planning	Views the tasks that are being planned.
   Starting	Views the tasks that start running.
   Running	Views running tasks.
   Finishing	Views the tasks that are being finished.
   Finished	Views the finished tasks.
   Failed	Views failed tasks, which can be filtered by failure cause.

5. Click a task ID to view the basic information, resource usage, stages, and tasks. For a failed task, you can view related logs on the detail query page.
   Figure 3 Viewing task details
   
   Figure 4 Task resource utilization summary
   
   Figure 5 Stages
   

   Table 3 Stages monitoring information

   Monitoring Item	Description
   SCHEDULED TIME SKEW	Indicates the scheduled time of the concurrent tasks on a node in the current stage.
   CPU TIME SKEW	Indicates whether concurrent tasks have computing skew in any stage phase.

   Figure 6 Tasks
   

   Table 4 Tasks monitoring items

   Monitoring Item	Description
   ID	Indicates the ID of the task that is concurrently executed in multiple phases. The format is Stage ID:Task ID.
   Host	Indicates the Worker node where the current task is being executed.
   State	Indicates the task execution status, including PLANNED, RUNNING, FINISHED, CANCELED, ABORTED, and FAILED.
   Rows	Indicates the total number of data records read by a task. The unit is thousand (k) or million (M). By analyzing the number of data records read by different tasks in the same stage, you can quickly determine whether data skew occurs in the current task.
   Rows/s	Indicates the number of data records read by a task per second. By analyzing the number of data records read by different tasks in the same stage, you can quickly determine whether the network bandwidth of the node is different and whether the node NIC is faulty.
   Bytes	Indicates the data volume read by a task.
   Bytes/s	Indicates the data volume read by a task per second.
   Elapsed	Indicates the task execution duration.
   CPU Time	Indicates the CPU time used by a task.
   Buffered	Indicates the buffer size of a task.

6. Click the "Host" link to view the task resource usage of each node.
   Figure 7 Resource usage of the Task node
   

   Table 5 Monitoring metrics of node resources

   Name	Description
   Node ID	Indicates the host ID.
   Heap Memory	Indicates the maximum heap memory size.
   Processors	Indicates the number of processors.
   Uptime	Indicates the running duration.
   External Address	Indicates the external IP address.
   Internal Address	Indicates the internal IP address.
   Process CPU Utilization	Indicates the physical CPU utilization.
   System CPU Utilization	Indicates the system CPU utilization.
   Heap Utilization	Indicates the heap memory utilization.
   Non-Heap Memory Used	Indicates the non-Heap memory size.
   Memory Pools	Indicates the memory pool size of the current Worker node.