Configure alert rules

Prerequisites

• An ApsaraDB for SelectDB service-linked role, AliyunServiceRoleForSelectDB, must be created. By default, this role has permission to access the ARMS service. For more information, see Service-linked role.

• If you want to report monitoring data from your ApsaraDB for SelectDB instance to your own ARMS for centralized monitoring and alerting, you must first activate the ARMS service. For more information, see Activate ARMS.

Procedure

1. Log on to the ApsaraDB for SelectDB console.

2. In the upper-left corner of the page, select the region where the instance is located.

3. On the Instances page, click the target Instance ID to go to the Instance Details page.

4. In the left-side navigation pane, click Monitoring and Alerts.

5. On the Monitoring and Alerts page, click the Alert Management tab, and then click Create SelectDB Monitoring Alert.

   Note

   After you click the Alert Management tab, the SelectDB Monitoring Alert List page may take 3 to 5 seconds to load.

6. On the Create SelectDB Monitoring Alert page, configure the parameters.

   You can create an alert rule using either a static threshold or a custom PromQL query.

   • Static threshold: Provides preset alert metrics. You can select a metric to quickly create an alert rule by defining a condition.

   • Custom PromQL: Create an alert rule by writing a custom PromQL query. This is useful for metrics that are not in the preset list.

   Static threshold

   Parameter	Description	Example
   Alert Rule Name	The name of the alert rule.	CPU utilization alert
   Check Type	Select Static threshold.	static threshold
   Instance	Select the instance for the alert rule. The default value is Traverse, which means the alert rule applies to all instances.	selectdb-cn-7213n****
   Cluster	Select the cluster for the alert rule. The default value is Traverse, which means the alert rule applies to all clusters.	selectdb-cn-7213n****-be
   Alert contact group	Select an alert contact group. The supported alert contact groups vary based on the Prometheus instance type. The available options change based on the selected Prometheus instance type.	SelectDB monitoring alert
   Alert metric	Select the metric for the alert. Each alert contact group provides a different set of metrics.	CPU utilization
   Alert condition	Define the condition that triggers an alert based on the selected metric.	An alert is sent when the CPU utilization is greater than 80%.
   Filter Conditions	No filter	None
   Data Preview	The Data Preview section displays the PromQL query corresponding to the alert condition and charts the metric's values over time. By default, the chart shows the real-time value for a single resource. Use the filters in this section to view data for different resources and time ranges. Note The alert threshold is displayed as a red dashed line in the chart. The curve turns red when the threshold is met and remains blue otherwise. Hover over the time-series curve to view resource details at a specific point in time. Select a time range on the time series chart to view the curve for that period.	None
   Duration	Generate alert event when condition is met: Triggers an alert if any data point meets the threshold. Generate alert event when condition persists for N minutes: Triggers an alert only if the condition persists for at least N minutes.	1
   Alert Level	The severity level of the alert. The default level is Default. Severity increases in the order: Default, P4, P3, P2, and P1.	P2
   Alert Content	The content of the alert notification that users receive. You can use Go template syntax to customize parameter variables in the alert message.	Node: {{$labels.pod_name}} CPU utilization {{$labels.metrics_params_opt_label_value}} {{$labels.metrics_params_value}}%, current value {{ printf "%.2f" $value }}%
   Alert Notification	Simple Mode: Lets you quickly configure the Notification Contacts, Notification Period, and Whether to Resend Notifications. Standard Mode: Lets you select a notification policy. If no policy exists, click Create Notification Policy to create one. For more information, see Notification policy. Important Quickly specifying a notification policy ensures that alert events from the current alert rule are matched by the selected notification policy. However, these events may also be matched by other notification policies that use fuzzy matching. The relationship between alert events and notification policies is many-to-many.	None
   Advanced Settings	Alert Check Cycle: The interval at which the alert rule is evaluated. The default and minimum value is 1 minute. Check After Data Is Complete: Ensures data is complete before evaluation. This is enabled by default, and we recommend keeping this setting. Tags: Assign tags to the alert. Notification policies can use these tags for matching. Annotations: Set annotations for the alert.	Alert check cycle: 1 minute Check After Data Is Complete: Yes Tags (labels): None Annotations: None

   Custom PromQL

   Parameter	Description	Example
   Alert Rule Name	The name of the alert rule.	Pod CPU utilization greater than 80%
   Check Type	Set to Custom PromQL.	Custom PromQL
   Instance	Select the instance for the alert rule.	selectdb-cn-7213n****
   Cluster	Select the cluster for the alert rule.	selectdb-cn-7213n****-be
   Reference Alert Contact Group	Select an alert contact group. The supported alert contact groups vary based on the Prometheus instance type. The available options change based on the selected Prometheus instance type.	SelectDB monitoring alert
   Reference Metrics	(Optional) Select a reference metric to use its PromQL template as a starting point. You can then modify the pre-filled query. The available reference metrics are automatically filtered based on the selected Prometheus instance type. Note The template provided by the Reference Metrics is incomplete. You must modify it to create a valid Custom PromQL Statements.	99th Percentile Query Latency
   Custom PromQL Statements	Use a PromQL query to define the alert expression.	avg(doris_fe_query_latency_ms{quantile="0.99",pod=~,cluster_id=~}) by (cluster_id) > 300
   Data Preview	The Data Preview section displays the PromQL query corresponding to the alert condition and charts the metric's values over time. By default, the chart shows the real-time value for a single resource. Use the filters in this section to view data for different resources and time ranges. Note The alert threshold is displayed as a red dashed line in the chart. The curve turns red when the threshold is met and remains blue otherwise. Hover over the time-series curve to view resource details at a specific point in time. Select a time range on the chart to zoom in on the time-series curve for that period.	None
   Duration	If the alert condition is met, an alert event is generated.: Triggers an alert if any data point meets the threshold. Generate alert event when condition persists for N minutes: Triggers an alert only if the condition persists for at least N minutes.	1
   Alert Level	The severity level of the alert. The default level is Default. Severity increases in the order: Default, P4, P3, P2, and P1.	Default
   Alert Content	The content of the alert notification that users receive. You can use Go template syntax to customize parameter variables in the alert message.	Namespace: {{$labels.namespace}}/Pod: {{$labels.pod_name}}/Disk device: {{$labels.device}} utilization exceeds 90%, current value {{ printf "%.2f" $value }}%
   Alert Notification	Simple Mode: Lets you quickly configure the Notification Contacts, Notification Period, and Whether to Resend Notifications. Standard Mode: Lets you select a notification policy. If no policy exists, click Create Notification Policy to create one. For more information, see Notification policy. Important Quickly specifying a notification policy ensures that alert events from the current alert rule are matched by the selected notification policy. However, these events may also be matched by other notification policies that use fuzzy matching. The relationship between alert events and notification policies is many-to-many.	None
   Advanced Settings	Alert Check Cycle: The interval at which the alert rule is evaluated. The default and minimum value is 1 minute. Check After Data Is Complete: Ensures data is complete before evaluation. This is enabled by default, and we recommend keeping this setting. Tags: Assign tags to the alert. Notification policies can use these tags for matching. Annotations: Set annotations for the alert.	Alert check cycle: 1 minute Check After Data Is Complete: Yes Tags (labels): None Annotations: None

7. After you configure the parameters, click Save. The alert rule becomes active immediately.

Recommended alert configurations

You can configure alerts based on your requirements. The following table lists common metrics and provides recommended settings.