Set Up Alarm Definitions for HeatWave DB Systems

You can set up alarm definitions to simplify the process of creating and managing alarms for HeatWave DB systems in Database Management.

In the Alarm definitions section, you can use the available standardized set of rules and thresholds and create Oracle-recommended alarms to proactively monitor important metrics such as CPU and disk space utilization. Oracle-recommended alarms in Database Management greatly simplify the process of creating an Oracle Cloud Infrastructure Monitoring service alarm and setting up a notification system for common operational scenarios for your DB systems. This process notifies you of health and performance issues and serves as a proactive monitoring mechanism to ensure that the DB system operates efficiently and securely. In addition, if you want to use the complete set of options available when creating alarms, you have the option of navigating to the Monitoring service and creating custom alarms for the DB system. Here are a few other benefits of the Alarm definitions feature:

  • Edit and customize alarms: You can edit the preconfigured values for Oracle-recommended alarms and the values for custom alarms to customize them. This provides you with the flexibility to edit an alarm to meet specific requirements.
  • Clone alarms for multiple DB systems: You can clone previously created alarms and apply them to specific DB systems or all the DB systems in a compartment, thereby ensuring standardized monitoring settings.

To go to the Alarm definitions section, go to the MySQL database details page and click Alarm definitions on the left pane under Resources.

In the Alarm definitions section, you can:

  • Create Oracle-recommended alarms or navigate to the Monitoring service to create custom alarms for a DB system.
  • View the alarms created for the DB system or all the DB systems in the compartment.
    Note

    To view alarms in the Alarm definitions section, you must ensure that:
    • The alarms are created using the Oracle Cloud Infrastructure (out-of-the-box) metrics emitted in the oracle_oci_database (resource group: mysql_database) and oci_mysql_database namespaces.
    • The resourceID dimension is specified when creating alarms for a DB system. The alarms created for all the DB systems in a compartment do not require dimensions to be specified.

    For information on:

  • Edit alarms
  • Clone alarms
  • Delete alarms

In addition to Database Management permissions, other Oracle Cloud Infrastructure service permissions are required to perform alarm-related tasks. For information on:

  • Monitoring service permissions required to create alarms, see Managing Alarms.
  • Oracle Cloud Infrastructure Notification service permissions required to create and use topics, see Managing Topics.

Create Recommended Alarms

You can create Oracle-recommended alarms that are preconfigured for common operational scenarios for your DB systems.

  1. In the Alarm definitions section on the MySQL database details page, click Create and then click Recommended alarms.
    Note

    If you want to create a custom alarm for your DB system in the Oracle Cloud Infrastructure Monitoring service instead of an Oracle-recommended alarm, click Create and then click Custom alarm. To view a custom alarm in the Alarm definitions section in Database Management, you must ensure that:
    • The alarm is created using an out-of-the-box metric emitted in the oracle_oci_database (resource group: mysql_database) and oci_mysql_database namespaces.
    • The resourceID dimension is specified when creating an alarm for a DB system. An alarm created for all the DB systems in a compartment does not require dimensions to be specified.

    For information on how to create an alarm in the Monitoring service, see Creating a Basic Alarm.

  2. In the Create recommended alarms panel, select from the available, commonly used alarms. By default, all the alarms are selected.

    For information on the Oracle-recommended alarms available for HeatWave DB systems, see Oracle-recommended Alarms for HeatWave DB Systems.

  3. Optionally, click the Actions icon (Actions) available for each recommended alarm and click Edit threshold to edit the preconfigured values for the alarm. For example, the CPU utilization (%) alarm is fired when the average CPU utilization is greater than 90% over the last 5 minutes. In the Edit alarm panel, you can make changes to the various values pertaining to the alarm, for example, specify a different threshold value, interval (time period) value for querying metric data, or the trigger delay period, which is the time period before the alarm is in firing state. For information on the fields displayed in the Edit alarm panel, see Edit Alarms.
  4. In the Notifications section, select the compartment and topic to which you want to send alarm notifications.
    If an existing topic is not available, click Create a topic to create a new topic and specify subscription details. For information on topics and subscriptions, see Managing Topics and Managing Subscriptions.
  5. Click Create.
    The Create alarms panel is displayed with a status of the Oracle-recommended alarms being created.
  6. Close the Create alarms panel after the alarms are created.

Once the alarms are created, you can:

  • Click the alarm to view the alarm details on the Alarm Definitions page in the Monitoring service. For Oracle-recommended alarms, the provider: DBM free-form tag is added by default during the creation process and is displayed in the Tags section.
    Note

    It's possible that different users create the same Oracle-recommended alarm for a DB system. If duplicate alarms are displayed for a DB system, click the alarm to view information such as who created the alarm and when it was created in the Tags section on the Alarm Definitions page in the Monitoring service.
  • Use the Search field and the filter drop-down lists above the list of alarms to filter the alarms:
    • Severity filter: Select an alarm severity option to only view the alarms of a particular severity type.
    • Scope filter: Select one of the available options to view the alarms specified for all the DB systems in a compartment or the alarms specified for the DB system. By default, all the alarms are displayed.
    • Origin filter: Select one of the available options to view the Oracle-recommended alarms or the custom alarms created in the Monitoring service, for the DB system. By default, all the alarms are displayed.
  • Click the Actions icon (Actions) and click Edit threshold to edit the basic alarm values in Database Management or click Edit alarm to view and edit the complete set of values specified for the alarm in the Monitoring service. For information, see Edit Alarms.
  • Select alarms and click Clone to clone the alarms. For information, see Clone Alarms.
  • Select alarms and click Delete to delete the alarms.

Oracle-recommended Alarms for HeatWave DB Systems

Here's the list of the Oracle-recommended alarms available for HeatWave DB systems.

The Monitoring stopped - Metric collection issue Oracle-recommended alarm uses the MonitoringStatus metric in the oracle_oci_database namespace (resource group: mysql_database) and the remaining Oracle-recommended alarms for HeatWave DB systems use metrics in the oci_mysql_database namespace.

Oracle-recommended Alarm Severity Description
CPU utilization (%) Critical Alarm created using the CPUUtilization metric to notify when the CPU utilization for the DB system host or HeatWave nodes exceeds the specified threshold.

By default, this alarm is fired when the mean of the CPUUtilization metric is greater than the threshold value of 90%.

Default alarm query: CPUUtilization[5m].mean() > 90

CPU utilization (%) Warning Alarm created using the CPUUtilization metric to notify when the CPU utilization for the DB system host or HeatWave nodes exceeds the specified threshold.

By default, this alarm is fired when the mean of the CPUUtilization metric is greater than the threshold value of 75%.

Default alarm query: CPUUtilization[5m].mean() > 75

Disk space utilization (%) Critical Alarm created using the DbVolumeUtilization metric to notify when the total space utilization of the DB system volumes exceeds the specified threshold.

By default, this alarm is fired when the mean of the DbVolumeUtilization metric is greater than the threshold value of 90%.

Default alarm query: DbVolumeUtilization[5m].mean() > 90

Disk space utilization (%) Warning Alarm created using the DbVolumeUtilization metric to notify when the total space utilization of the DB system volumes exceeds the specified threshold.

By default, this alarm is fired when the mean of the DbVolumeUtilization metric is greater than the threshold value of 80%.

Default alarm query: DbVolumeUtilization[5m].mean() > 80

Disk write bytes Critical Alarm created using the DbVolumeWriteBytes metric to notify when the total bytes written to the DB system volumes is equal to 0.

By default, this alarm is fired when the mean of the DbVolumeWriteBytes metric is 0.

Default alarm query: DbVolumeWriteBytes[5m].mean() == 0

HeatWave health status Critical Alarm created using the HeatWaveHealth metric to notify when the HeatWave cluster health status is not healthy, and exceeds the specified threshold.

By default, this alarm is fired when the mean of the HeatWaveHealth metric is greater than 0.

Default alarm query: HeatWaveHealth[5m].mean() > 0

Memory utilization (%) Critical Alarm created using the MemoryUtilization metric to notify when the memory utilization for the DB system host or HeatWave nodes exceeds the specified threshold.

By default, this alarm is fired when the mean of the MemoryUtilization metric is greater than the threshold value of 95%.

Default alarm query: MemoryUtilization[5m].mean() > 95

Memory utilization (%) Warning Alarm created using the MemoryUtilization metric to notify when the memory utilization for the DB system host or HeatWave nodes exceeds the specified threshold.

By default, this alarm is fired when the mean of the MemoryUtilization metric is greater than the threshold value of 90%.

Default alarm query: MemoryUtilization[5m].mean() > 90

Monitoring stopped - Metric collection issue Critical Alarm created using the MonitoringStatus metric to notify when the DB system is down. This alarm is fired when the metric collection has stopped due to issues related to network, agent, credentials, connection, or missing data or when the DB system is down.

This is an advanced alarm and cannot be edited in Database Management. You can edit this alarm in the Oracle Cloud Infrastructure Monitoring service, however, it's recommended that you do not edit it.

Network receive bytes Critical Alarm created using the NetworkReceiveBytes metric to notify when the total network receive bytes for the DB system is equal to 0.

By default, this alarm is fired when the mean of the NetworkReceiveBytes metric is 0.

Default alarm query: NetworkReceiveBytes[5m].mean() == 0

Edit Alarms

You can edit the values specified for an alarm to customize the alarm to meet your requirements.

The editing capability in the Alarm definitions section provides you with the flexibility to make changes to alarm values and customize the alarm. In this section, you can:

  • Click Edit threshold when creating an Oracle-recommended alarm or for a previously created Oracle-recommended or custom alarm to edit basic alarm values such as threshold, interval, and severity. For Oracle-recommended alarms, this enables you to make changes to the preconfigured values to meet specific requirements.
  • Click Edit alarm for a previously created Oracle-recommended or custom alarm to view and edit the complete set of values specified for the alarm in the Oracle Cloud Infrastructure Monitoring service.

To edit basic alarm values for Oracle-recommended or custom alarms in Database Management:

  1. In the Alarm definitions section on the MySQL database details page, click the Actions icon (Actions) for an alarm and click Edit threshold.
    Note

    To edit the complete set of values specified for the alarm in the Monitoring service, click the Actions icon (Actions) for the alarm and click Edit alarm. To view and edit a custom alarm in the Alarm definitions section in Database Management, you must ensure that:
    • The alarm is created using an out-of-the-box metric emitted in the oracle_oci_database (resource group: mysql_database) and oci_mysql_database namespaces.
    • The resourceID dimension is specified when creating an alarm for a DB system. An alarm created for all the DB systems in a compartment does not require dimensions to be specified.
  2. In the Edit alarm panel, edit the following details, as required:
    1. Name: Name of the alarm.
    2. Alarm body: Human-readable content of the notification.
    3. Severity: Type of response required when the alarm is in the firing state. For example, Critical, Error, and so on.
    4. Interval: Aggregation window, or the frequency at which data points are aggregated.
    5. Statistic: Function to use to aggregate the data points. For example, Mean, Rate, and so on.
    6. Operator: Operator used in the condition threshold. For example, greater than, equal to, and so on.
    7. Threshold: Value to use for the condition threshold. For example, for the CPU utilization (%) Oracle-recommended alarm, the default threshold value is 90 and you can change the value to 80 to fire the alarm when the mean of the CpuUtilization metric is greater than 80.
    8. Trigger delay minutes: Number of minutes that the condition must be maintained before the alarm is in the firing state.
    9. Aggregate metric streams: Option to return the combined value of all metric streams for the selected statistic. If editing an Oracle-recommended alarm, note that this check box is deselected by default.
    10. Split notifications per metric stream: Option to individually track metric status by metric stream and send a message when metric status for each metric stream changes. For more information on this option, see Scenario: Split Messages by Metric Stream.
  3. Click Save.

Clone Alarms

You can clone existing Oracle-recommended or custom alarms and apply them to specific DB systems or all the DB systems in a compartment.

The cloning capability reduces the effort required to create multiple alarms, and allows you to standardize alarm settings across multiple DB systems.

To clone alarms:

  1. In the Alarm definitions section on the MySQL database details page, select one or more alarms and click Clone.
    Note

    When cloning multiple alarms at the same time, you must ensure that you select alarms with the same scope.
  2. In the Clone alarms panel, select one of the following options:
    • Clone to compartment: Select this option and then a compartment in the Select compartment drop-down list to clone the alarms for all the DB systems in the compartment.
    • Clone to selected database: Select this option and a compartment in the Select compartment drop-down list, and then select the DB systems to clone the alarms and apply them to the selected DB systems.
  3. Click Clone.
Note

When cloning alarms, you may create duplicate alarms. If duplicate alarms are listed for the DB systems, click the alarm to view information such as who created the alarm and when it was created in the Tags section on the Alarm Definitions page in the Oracle Cloud Infrastructure Monitoring service.