Configuring clustersMonitoring and alertsenables users to monitor the operational status and resource status of clusters in real time, promptly detect and handle anomalies, and ensure business stability. This article provides several types of basic Monitoring and alertsconfigurations and recommends applying these basic configurations.

1. Log in to the Cloud Monitoring Console . On the Alert center>Alert Template>Preset Alert Template page, under Resource Type , select Elastic Compute>Cloud Server to filter templates. For detailed instructions on creating and using alert templates, see Create Alert Template .
   
2. Click the Cloud Server Preset Alert Template name to enter the template details page and view information such as alert trigger conditions, alert levels, and associated alert policies.
   
3. On the right side of the Cloud Server Preset Alert Template details page, click Enable, or return to the Preset Alert Template page and click the Apply with One Click button. The system will automatically generate and enable alert policies, applying them to all cloud server resources under this Volcano Engine account.

   tip

   After disabling the alert template, all alert policies generated by the preset alert template will be deleted accordingly.

On the Alert center>Alert Template>Preset Alert Template page, in Resource Type , select Data Platform>E-MapReduce to filter the templates. Refer to Configure cloud server resource alertsconfiguration steps to enable the E-MapReduce preset alert template.

1. Log in to the Cloud Monitoring Console and go to the Event Center>Event Rules>Create Event Rule page.
2. Create service status/operation exception event alert rules,
   1. In the Resource Type dropdown, select E-MapReduce.
   2. In the Event Name dropdown, enter the keyword "status" to filter service status events and select event names. It is recommended to select all status events for services in use.
   3. In the Event Name dropdown, enter keywords such as "success" or "failure" for operation event keywords and select event names. You can configure based on high-frequency operations.
      For detailed guidance on creating event rules, see Create Event Rule .
      

Core service metrics directly affect business stability. It is recommended to configure alert policies for these metrics to promptly detect and address anomalies.

1. Log in to the EMR console, go to the EMR cluster details page>Alert center>Subscribe to event alerts, and navigate to the Cloud Monitoring alert policy page.
   

2. On the Cloud Monitoring alert policy page, click Create alert policy to enter the creation page.

3. Set Resource type to Data Middle Platform>E-MapReduce.
   
   

4. Configure alert Dimension, metric types Trigger condition, and other alert policy details. For detailed parameter descriptions, see Create Alert Policy .
   The table below lists key service metrics and recommended alert policies. Please refer to these policies for configuration.

   Alert dimension	Alert metric and Note	Metric level	Trigger condition
   HDFS	DFS space used by HDFS	Critical	Alert when usage exceeds 90% of total capacity Total capacity query path: EMR Console > Cluster Details > Cluster Monitoring > HDFS > Overview > Storage Status. tip Recommendation: Increase the JVM heap size of the NameNode and clean up directories containing a large number of small files Alert sending interval recommendation: Check every 5 minutes; if it occurs twice within 30 minutes, trigger an alert.
   	HDFS storage utilization	Critical	Alert when above 80%
   	Number of lost HDFS blocks	Critical	Configure as needed. If lost blocks are not allowed, set alert for Missing block > 0.
   	NameNode heap space usage ratio	Critical	Alert when above 80%
   	DataNode old generation space utilization	Critical	Alert when above 95% tip Alert sending interval recommendation: The alert should persist for a certain period within a specified time frame.
   Yarn	ResourceManager old generation space	Critical	Alert when above 95%
   	NodeManager old generation space	Critical	Alert when above 95%
   Hive	Number of open hiveserver2 connections	Critical	Set as needed
   	HiveMetaStore heap memory utilization	Critical	Alert when above 95%
   	HiveServer2 heap memory utilization		Alert when above 95%
   	Number of HiveServer2 query failures		Set as needed
   Hbase	HRegion full gc	P2	Monitor and troubleshoot if it occurs frequently.
   	HRegion read rate	P2	Set as needed
   	HRegion write rate	P2	Set as needed
   Presto/Trino Coordinator	Current amount of HeapMemory used by JVM	P1	Alert for exceeding 95% of heap memory
   Presto/Trino Worker	Current amount of HeapMemory used by JVM	P1	Alert for exceeding 95% of heap memory
   ZooKeeper	Number of ZooKeeper znodes	P2	Set as needed
   ZookeeperOverview	Total number of connections in ZooKeeper cluster	P2	Set as needed
   Kafka（KafkaConsumerGroup）	Kafka Consumer Group Lag	P1	Set as needed