Interconnecting MRS with Cloud Eye
When an MRS cluster is running, administrators need to continuously monitor the status of cluster components and node resource usage (such as CPU, memory, and disk usage). While MRS Manager provides real-time monitoring within the cluster, it lacks unified cross-cluster monitoring, long-term trend analysis, and alarm notification.
Huawei Cloud Eye is a unified monitoring platform for cloud resources. It can interconnect with MRS. Once metric sharing is enabled, MRS pushes key monitoring metrics of cluster components and nodes to Cloud Eye. You can then view monitoring trends, configure alarm rules, and receive alarm notifications via Cloud Eye console. Additionally, monitoring data can be retrieved via the Cloud Eye API to integrate with the self-built O&M platforms. This document is intended for MRS cluster administrators and O&M engineers.
Typical application scenarios:
- Unified cross-cluster monitoring: View the monitoring data of multiple MRS clusters centrally on the Cloud Eye console without logging in to Manager of each cluster.
- Alarm notification configuration: Set alarm thresholds for key metrics (such as node CPU usage and HDFS disk usage) of MRS clusters in Cloud Eye and receive alarm notifications by SMS or email.
- Long-term trend analysis: Leverage Cloud Eye's long-term data storage capability to analyze the historical monitoring trend of MRS clusters, aiding in capacity planning and performance optimization.
Prerequisites
The user enabling the metric sharing function have been granted the mrs:ops:grant or mrs:cluster:updateOps permission.
Enabling Metric Sharing
- Log in to the MRS console.
- In the active MRS cluster list, click the name of the target cluster to go to its Dashboard page.
- In the upper right corner of the page, choose O&M > Share Metric to enable metric sharing.
Verification method:
- If the Disable Metric Sharing option is displayed in the O&M menu in the upper right corner of the page, metric sharing has been successfully enabled.
- A while after the metric sharing is enabled, the monitoring data of the MRS cluster will be pushed to Cloud Eye.
- Log in to the Cloud Eye console.
- In the navigation pane on the left, choose Cloud Service Monitoring. On the displayed page, click MapReduce Service to view the metrics of the MRS cluster interconnected with Cloud Eye.
Follow-up procedure
- On the MRS monitoring page, you can view the real-time values and historical trends of each metric.
- To configure an alarm rule, choose Alarm Rules > Create Alarm Rule, select an MRS cluster and metric, and set the alarm threshold and notification mode.
Disabling Metric Sharing
- Log in to the MRS console. In the active cluster list, click the name of the target cluster to go to its Dashboard page.
- In the upper right corner of the page, choose O&M > Disable Metric Sharing to disable metric sharing.
Verification method:
If the Share Metric option is displayed in the O&M menu in the upper right corner of the page, metric sharing has been successfully disabled. After metric sharing is disabled, MRS stops pushing new monitoring data to Cloud Eye.
Helpful Links
- For details about the metrics that can be monitored for an MRS cluster, see Basic MRS Monitoring Metrics.
- For details about how to configure and view alarms for an MRS cluster on Manager, see Viewing Alarms of an MRS Cluster.
FAQs
Q: Does metric sharing consume MRS cluster resources?
A: Metric sharing consumes only few cluster resources. MRS periodically pushes collected monitoring metric data to Cloud Eye, which does not increase the cluster load. The consumed resources are released after the metric sharing is disabled.
Q: Do the alarm rules configured for MRS clusters in Cloud Eye remain effective after metric sharing is disabled?
A: No. Since MRS stops pushing monitoring data to Cloud Eye, Cloud Eye cannot retrieve the latest metric values, and related alarm rules cannot be triggered. After metric sharing is enabled again, the alarm rules will automatically resume effectiveness.
