Updated on 2023-11-03 GMT+08:00

Elasticsearch Monitoring

Introduction

An Elasticsearch monitor detects your database status, and automatically triggers traffic switching when the database is abnormal.

Table 1 Monitoring status description

Monitoring Status

Description

Green

The monitoring is normal.

Red

The monitoring is abnormal. Traffic switchover is automatically triggered (if enabled) when the status changes from normal to failed.

Yellow

The monitor initialization failed. The initial detection fails when the monitor starts. In this case, confirm the configurations or contact O&M personnel.

Light gray

Monitor creation or configuration failed

Dark gray

Invocation failed. The ETCD connection status is abnormal.

Creating an Elasticsearch Monitor

  1. Log in to the MAS console. On the Multi-Active Instances page, click an instance to go to its console.
  2. Click the Monitor List tab, and click Create Monitor.
  3. Configure the basic information, then click Next: Data Centers.

    Figure 1 Basic information configurations
    Table 2 Basic information parameters

    Parameter

    Description

    Monitor

    Select Elasticsearch Monitoring.

    Application

    Select the application.

    Monitor Name

    Customize the monitor name.

    Exception Notification

    By default, this option is disabled.

    If it is enabled,

    • monitor and database exceptions will be sent to you in a timely manner with the Huawei Cloud SMN service. Configure a secret key first.
    • Cloud Eye is enabled by default. Ensure that the alarm rules for MAS have been created on the Cloud Eye console. For details, see Creating an Alarm Rule to Monitor an Event.

    Subject

    If Exception Notification is enabled, select a subject from the drop-down list or click Add to create a new one.

    Monitoring

    The default value is Yes. If No, database exceptions will not be monitored.

    Automatic Switchover

    The default value is Yes. If No, automatic switchover of the databases will not be triggered.

    Routing Algorithm

    Option: Single read/write

  4. Configure the data centers, then click Next: Advanced.

    Figure 2 Data center configurations
    Table 3 Data center parameters

    Parameter

    Description

    Cloud

    The environment where the monitored Elasticsearch database is deployed

    Region

    The region where the monitored Elasticsearch database is deployed

    AZs

    Set this parameter for data centers 1 and 2 based on the AZs where the Elasticsearch databases are deployed.

    Connection Address

    The IP address and port number for accessing the Elasticsearch database

    Add Connection Address

    You can add more connection addresses.

    Username

    Enter the username for logging in to the Elasticsearch database.

    CAUTION:

    To prevent risks, it is recommended to use a username that is different from your service username and has been granted only the read permissions.

    Protocol

    Options: HTTPS and HTTP

    Password

    Enter the password of the monitored Elasticsearch database.

    Confirm Password

    Enter the password again.

    MAS supports Elasticsearch 7.10.2 and later. Ensure that the password is correct, otherwise, the data source will be locked for 15 minutes.

  5. Configure the advanced settings, then click Next: Confirm.

    Figure 3 Advanced configurations
    Table 4 Advanced settings

    Parameter

    Description

    Retry Interval (ms)

    The interval of reconnection attempts, in milliseconds. Default value (recommended): 3000

    Value range: 1 to 300,000

    Monitoring Timeout (ms)

    The timeout duration before a monitor becomes abnormal, in milliseconds.

    Default value (recommended): 40,000; Value range: 1 to 600,000

    NOTE:

    Ensure that the time you specify here is longer than the time required for data center 1 to synchronize data to data center 2. Otherwise, the data in data center 2 may be incomplete.

  6. Confirm settings and click Create.

    If the created monitor is not normal, that is, its indicator is not green, its configurations or databases may be abnormal. In this case, rectify the fault.