You can create alerts to define alert rules for specific monitored objects. When an alert rule is triggered, the system sends an alert notification to the specified contact group based on the specified notification method. This reminds you to take necessary actions to solve the problem.

Background information

By default, alert notifications are sent based on the following rules:

  • To prevent you from receiving a large number of alert notifications in a short period of time, the system sends only one message for repeated alerts within 24 hours.
  • If no repeated alerts are generated within 5 minutes, the system sends a recovery email to notify you that the alert is cleared.
  • After a recovery email is sent, the system resets the alert status. If this alert arises again, it is deemed a new one.

An alert widget is essentially a data display method for datasets. When an alert widget is created, a dataset is also created to store the underlying data of the alert widget.

Note New alerts take effect within 10 minutes. The alert check may have a delay of 1 to 3 minutes.

To create an alert for an application monitoring job on Java Virtual Machine-Garbage Collection (JVM-GC) times in period-over-period comparison, perform the following steps:

  1. Log on to the Enterprise Distributed Application Service (EDAS) console.
  2. In the left-side navigation pane, click Applications. In the top navigation bar, select the region where the application whose alert rules you want to manage is deployed. In the upper part of the Applications page, select the namespace where the application is deployed.
  3. In the Create Alarm dialog box, enter all required information and click Save.
    1. Specify Alarm Name. For example, you can enter alert on JVM-GC times in period-over-period comparison.
    2. From the Application Site drop-down list, select an application. From the Application Group drop-down list, select an application group.
    3. From the Type drop-down list, select a metric type. For example, you can select JVM_Monitoring.
    4. Set Dimension to Traverse.
    5. Specify Alarm Rules.
      1. Select Meet All of the Following Criteria.
      2. Edit an alert rule. For example, an alert is triggered when the value of N is 5 and the average value of JVM_FullGC increases by 100% compared with that in the previous hour.
        Note To create more alert rules, click the + icon on the right side of the created alert rule.
    6. Specify Notification Mode. For example, you can select Email.
    7. Specify Notification Receiver. In the Contact Groups list, click the name of a contact group. If the contact group appears in the Selected Groups list, the setting is successful.
    Application Monitoring Alarm

Basic fields for general use

The following table describes the basic fields in the Create Alarm dialog box.

Field Description Setting
Application Site The monitoring job that is created. This setting cannot be changed because the current application is already deployed in a specific region.
Type The type of the metric. The types include Invocation_Statistic, Invocation_Type, Database_Metric, JVM_Monitoring, Host_Monitoring, and Exception_Invocation.
Dimension The dimension for the specified metric (dataset). You can select None, =, or Traverse.
  • If Dimension is set to None, the alert content shows the sum of all values of this dimension.
  • If Dimension is set to =, you must enter a specific value.
  • If Dimension is set to Traverse, the alert content shows the dimension content that actually triggers the alert.
Last N Minutes The system checks whether the data results in the last N minutes meet the trigger condition. Valid values of N: 1 to 60.
Notification Mode Email, SMS, Ding Ding Robot, and Webhook are supported. You can select multiple modes. To configure a DingTalk chatbot alert, see .
Alarm Quiet Period You can turn on or off the Alarm Quiet Period switch. By default, the switch is turned on.
  • When the Alarm Quiet Period switch is on, if data remains in the triggered state, the second alert notification is sent 24 hours after the first alert is triggered. When data is recovered, the system sends a data recovery notification and clears the alert. If the data triggers the alert one more time, the system sends the alert notification again.
  • When the Alarm Quiet Period switch is off, if the alert is continuously triggered, the system sends the alert notification every minute.
Alarm Severity Valid values include Warn, Error, and Fatal. None.
Notification Time The time range during which alert notifications can be sent. Out of the specified time range, the system does not send alert notifications but still records alert events. For more information about alert event records, see .
Notification Content The custom content of the alert notification. You can edit the default template. In the template, the following four variables must be specified: $AlarmName, $AlarmFilter, $AlarmTime, and $AlarmContent. The rest of the content can be customized. Other variables are not supported.

Complex fields for general use: period-over-period comparison

  • N-minute-on-N-minute comparison: Assume that β is the data in the last N minutes, and α is the data generated between the Nth minute and the 2Nth minute. The data can be the average, sum, maximum, or minimum of the specified dataset. The N-minute-on-N-minute comparison is the percentage increase or decrease when β is compared with α.
  • N-minute-on-N-minute hourly comparison: Assume that β is the data in the last N minutes, and α is the data in the last N minutes of the previous hour. The data can be the average, sum, maximum, or minimum of the specified dataset. The N-minute-on-N-minute hourly comparison is the percentage increase or decrease when β is compared with α.
  • N-minute-on-N-minute daily comparison: Assume that β is the data in the last N minutes, and α is the data in the last N minutes prior to the same time of the previous day. The data can be the average, sum, maximum, or minimum of the specified dataset. The N-minute-on-N-minute daily comparison is the percentage increase or decrease when β is compared with α.

Complex fields for general use: Alarm Data Revision

Valid values of the Alarm Data Revision field include Set 0, Set 1, or Set Null (Won't Trigger). The default value is Set Null (Won't Trigger). In most cases, this feature is used to fix data anomalies, including no data, abnormal compound indicators, and abnormal period-over-period comparisons.

  • Set 0: fixes the checked value to 0.
  • Set 1: fixes the checked value to 1.
  • Set Null (Won't Trigger): triggers no alert.

Scenarios:

  • Anomaly 1: no data

    User A wants to use the alert feature to monitor page views. When User A creates an alert, User A selects Browser Monitoring Alarm. User A sets the following alert rule: N is 5 and the sum of page views is no greater than 10. If the page is not accessed, no data is reported and no alert is sent. To solve this issue, you can select Set 0 for Alarm Data Revision. If no data is received, the system determines that zero data records are received. This meets the conditions specified in the alert rule. Therefore, an alert is sent.

  • Anomaly 2: abnormal compound indicators

    User B wants to use the alert feature to monitor the real-time unit price of a product. When User B creates an alert, User B selects Custom Monitoring Alarm. User B sets the dataset of variable a to the current total price, and the dataset of variable b to the current total number of items. User B also sets an alert rule. In this rule, N is 3 and the minimum value of the current total price divided by the current total number of items is less than or equal to 10. If the current total number of items is 0, the value of the compound indicator (the current total price divided by the current total number of items) does not exist. No alert is sent. To solve this issue, you can select Set 0 for Alarm Data Revision. If the value of the compound indicator (the current total price divided by the current total number of items) does not exist, the system determines that the value is 0. This meets the conditions specified in the alert rule. Therefore, an alert is sent.

  • Anomaly 3: abnormal period-over-period comparisons

    User C wants to use the alert feature to monitor the CPU utilization of a node. When User C creates an alert, User C selects Application Monitoring Alarm. User C also sets the following alert rule: N is 3 and the average CPU utilization of the node decreases by 100% compared with that in the previous monitoring period. If the CPU of the user fails in the last N minutes, the system cannot obtain α. This means that the period-over-period comparison result does not exist. No alert is sent. To solve this issue, you can select Set 1 for Alarm Data Revision. If α is not obtained, the system determines that the period-over-period comparison result is a decrease of 100%. This meets the conditions specified in the alert rule. Therefore, an alert is sent.

What to do next

You can query and delete alert records in alert management.