Updated on 2023-04-28 GMT+08:00

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.

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.

    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 worker nodes 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 node of the current instance.

  4. Filter query tasks by State on the Query History page.

    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 1 Viewing task details
    Figure 2 Task resource utilization summary
    Figure 3 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 4 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 5 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.