Dynamic threshold-triggered alert rules automatically detect anomalies in metrics such as CPU utilization, request count, traffic, duration, and consumption. When resource metrics breach the dynamic boundaries, Cloud Monitor sends alert notifications so that you can handle exceptions promptly.
Prerequisites
Ensure that you have either activated CloudMonitor Basic with the pay-as-you-go billing method or enabled a free trial for the dynamic threshold feature in the Trial Center. For more information, see Pay-as-you-go or Free trial of dynamic thresholds.
Limits
Dynamic thresholds for a time-series metric are subject to the following limitations:
-
Data: The total amount of data generated in the last 168 hours (7 days) must be no less than the amount of data that is supposed to be generated in 48 hours.
NoteMetric data must meet the preceding requirements when it is previewed.
-
Time: Dynamic thresholds take effect only when the alert rules are saved for about 20 minutes.
NoteAfter dynamic thresholds are generated, the system checks whether the metric data meet the requirements and is updated every 20 minutes. If the metric data meet the requirements and is updated, new dynamic thresholds are generated.
Procedure
-
Log on to the Cloud Monitor console.
-
In the left-side navigation pane, choose Alerts > Alert Rules.
-
On the Alert Rules page, click Create Alert Rule.
-
In the Create Alert Rule panel, configure the parameters for the alert rule.
Example: Create a dynamic threshold alert rule for the maximum CPU utilization of an Elastic Compute Service (ECS) instance, aggregated every minute. Cloud Monitor sends a warning-level alert to all contacts in the Alarm_Group alert contact group when the metric stays above the upper boundary for three consecutive periods.
-
For Product, select Elastic Compute Service (ECS). Under Resource Range, click the Instance tab, and then select one or more ECS instances for Associated Resources.
-
Configure rule conditions.
-
Click Add Rule.
-
In the Add Rule Description panel, enter a Rule Name. Under Metric Type, click the Dynamic Threshold tab. For Metric, select Instance Dimension > (ECS)CPU utilization. Set Statistics to Maximum over a 1 Minute period. For Alert Condition, select above Upper Boundary. Set Alert Sensitivity to High and the Trigger Condition to for 3 consecutive periods (1 period = 1 minute).
Higher sensitivity produces more alerts. Three sensitivity levels are available:
-
High sensitivity: The boundaries are close to the actual values. Small fluctuations trigger alerts. Suitable for core metrics of critical business.
-
Medium sensitivity: The boundaries are moderately close to the actual values. A moderate number of alerts are generated. Suitable for common metrics of important business.
-
Low sensitivity: The boundaries are far from the actual values. Only large fluctuations trigger alerts. Suitable for common metrics of non-critical business.
-
-
Click OK.
-
-
For Alert Contact Group, select Alarm_Group. Keep the other parameters at their default values or leave them unset.
NoteFor more information about alert rule parameters, see Parameter descriptions for alert rules.
-
-
Click Confirm.
-
View the dynamic threshold-triggered alert rule you created.
On the Alert Rules page, select Dynamic Threshold from the Threshold Type filter. Find the target alert rule and click Details in the Actions column.
In the Alert Rule Details panel, the Details tab displays the dynamic threshold chart for the target instance in the Dynamic Threshold Graph section.