E-MapReduce:View node health status

Prerequisites

A cluster has been created. For more information, see Create a cluster.

Limitations

This feature is available only for DataLake, Dataflow, OLAP, DataServing, and Custom clusters.

View the latest node health status

1. Go to the node management page.

   1. Log on to the EMR console.

   2. In the top navigation bar, select the appropriate region and resource group.

   3. On the Clusters page, find your cluster and click Nodes in the Actions column.

2. On the node management page, a health summary is displayed for each node group.

   • Green number: The number of nodes in the Good state.

   • Yellow number: The number of nodes in the Warning state.

   • Red number: The number of nodes in the Abnormal state.

   • Gray number: The number of nodes in the Unknown and Stateless states.

   You can also click the icon for the target node group on the Node Management page. You can view the health status of each node in the Health Status column.

   A node can have one of five health statuses: Good, Warning, Abnormal, Unknown, and Stateless. Each status is represented by a unique icon.

   Icon	Health status	Description
   	Good	The node is operating normally.
   	Warning	The node is operating normally, but a health check item has detected a potential issue that requires attention.
   	Abnormal	The node is unavailable. A health check item has detected a critical issue that requires immediate action.
   	Stateless	The node is not being monitored, for example, during the installation process or after being manually stopped. No action is required.
   	Unknown	The health check results are unavailable. If your business workloads are running normally, no action is required.

View node health checks

1. On the Node Management page, click the icon for the target node group.

2. In the row for the target node, click View Check Items in the Health Status column.

3. In the panel that opens, you can view the latest health check results and the node's health check history.

   The following table describes the health check items. In the threshold definitions, u represents the measured value.

   Check item	Description	Threshold	Unit
   Node liveness	Checks whether the current node is alive and responsive.	None	-
   file descriptor usage	Checks the usage of file descriptors on the node.	Warning: 95 <= u < 99 Abnormal: u >= 99	%
   disk fault event	Checks for underlying disk fault events.	None	-
   System environment	Checks the availability of critical system components, such as important files, Java, and Python.	None	-
   Component service dependencies	Checks the availability of dependencies for cluster components, such as package files and storage directories.	None	-
   network transmit drop rate	Checks the outbound packet drop rate.	Warning: 1.0 <= u < 2.5 Abnormal: u >= 2.5	%
   network receive error rate	Checks the inbound packet error rate.	Warning: 0.1 <= u < 0.5 Abnormal: u >= 0.5	%
   disk I/O latency	Checks the average disk I/O latency.	Warning: 400 <= u < 800 Abnormal: u >= 800	ms
   network receive drop rate	Checks the inbound packet drop rate.	Warning: 1.0 <= u < 2.5 Abnormal: u >= 2.5	%
   network transmit error rate	Checks the outbound packet error rate.	Warning: 0.1 <= u < 0.5 Abnormal: u >= 0.5	%
   system fault event	Checks for underlying system fault events.	None	-
   CPU usage	Checks the CPU usage on the node.	Warning: 95 <= u < 99 Abnormal: u >= 99	%
   disk inode usage	Checks the disk inode usage.	Warning: 90 <= u < 99 Abnormal: u >= 99	%
   memory usage	Checks the memory usage on the node.	Warning: 95 <= u < 99 Abnormal: u >= 99	%
   disk space usage	Checks the disk space usage.	Warning: 90 <= u < 99 Abnormal: u >= 99	%