Updated on 2023-06-20 GMT+08:00

Configuring Cluster Monitoring

You can use Cloud Eye to monitor the created clusters. After configuring the cluster monitoring, you can log in to the Cloud Eye management console to view cluster metrics.

The procedure for configuring cluster monitoring:

  1. Creating Alarm Rules: Customize alarm rules for the monitoring metrics. Once a metric exceeds the threshold, the system will notify you by sending emails or HTTP/HTTPS requests.
  2. Configuring Monitoring Metrics: Configure monitoring metrics for a cluster or a node in the cluster.
  3. Viewing Monitoring Metrics: View the statistics of the monitoring metrics in specific periods.

Prerequisites

  • The cluster is in the Available or Processing status.
  • The cluster has been running properly for more than 10 minutes.

Recommended Monitoring Metrics

  • Cluster CPU and JVM usage. You are advised to configure the following monitoring metrics: average JVM heap usage, maximum JVM heap usage, average CPU usage, and maximum CPU usage.
  • Cluster write and query latency and throughput. You are advised to configure the following monitoring metrics: average index latency, average index rate, average search latency, and average QPS.
  • Cluster write and query queue and rejected tasks. You are advised to configure the following monitoring metrics: tasks in write queue, tasks in search queue, rejected tasks in write queue, and rejected tasks in search queue.

Creating Alarm Rules

  1. Log in to the Cloud Eye console.
  2. In the navigation pane on the left, choose Alarm Management > Alarm Rules.
  3. In the Resource Type column, select Cloud Search Service as criteria to search for alarm rules that meet the requirements.

    If no alarm rules are available, create one by referring to Creating an Alarm Rule and Notification. For details about how to set Resource Type and Dimension, see Table 1.

    Table 1 Alarm rule configuration parameter

    Parameter

    Description

    Remark

    Resource Type

    Type of the resource that the alarm rule is created for

    Select Cloud Search Service.

    Dimension

    Metric dimension of the selected resource type

    CSS supports two dimensions. Select a dimension as required.

    • CSS Clusters: Alarm rules are specified by cluster.
    • CSS Clusters - CSS Instances: Alarm rules are specified by node in a cluster.

Configuring Monitoring Metrics

  1. Create a monitoring panel. For details, see Creating a Monitoring Panel. If an available monitoring panel has been created, skip this step.
  2. Add CSS monitoring graphs. For details, see Adding a Graph.

    For details about how to set Resource Type and Dimension, see Table 2.

    Table 2 Graph configuration parameter

    Parameter

    Description

    Remark

    Resource Type

    Type of the resource to be monitored

    Select Cloud Search Service.

    Dimension

    Metric dimension

    CSS supports two dimensions. Select a dimension as required.

    • CSS Clusters: Monitoring is executed by cluster.
    • CSS Clusters - CSS Instances: Monitoring is executed by node in a cluster.

Viewing Monitoring Metrics

  1. Log in to the CSS management console.
  2. Choose Clusters. Locate the target cluster and choose More > View Metric in the Operation column.
  3. Select a time range.
  4. View the monitoring metrics.