Dataphin:Configure project monitoring and alert rules

Limits

• Monitoring and alerting can be configured only for Basic and Prod (production) projects.

• A maximum of 100 project-level monitoring and alert rules are supported.

Procedure

1. On the Dataphin home page, choose Develop > Task O&M from the top menu bar.

2. In the navigation pane on the left, choose Monitoring Management > Offline Task Monitoring.

3. On the Offline Task Monitoring page, click the Project Monitoring tab.

4. On the Project Monitoring tab, click + New Project Monitoring.

5. In the New Project Monitoring dialog box, configure the following parameters.

   	Parameter	Description
   Monitored Object	Rule Name	Enter a name for the rule. The name can be up to 36 characters long and can contain Chinese characters, letters, digits, underscores (_), and hyphens (-).
   	Project	Select the project to monitor. The drop-down list contains projects for which the current user has the O&M - Access Folder permission.
   	Monitoring Scope	Select All Tasks in Project or Filter by Task Type. All Tasks in Project: Monitors all tasks in the current project. Filter by Task Type: Filters tasks in the current project to monitor them by task type, which combines a task type and a scheduling type. Task Type: Includes Integration and Compute Tasks, Modeling Task, and All Task Types. If you set Task Type to Integration and Compute Tasks or Modeling Task, you can configure a sub-type. For the sub-type, you can select All or Custom. Modeling Task sub-types include Logical Dimension Table, Logical Fact Table, and Logical Summary Table. For Integration and Compute Tasks, the available sub-types vary depending on the compute engine. Refer to the options displayed on the page for details. Scheduling Type: Includes Recurring Task, One-Time Task, and All Scheduling Types. If you set Scheduling Type to Recurring Task, you can configure the scheduling cycle. For the scheduling cycle, you can select All or Custom. Scheduling cycles include year, month, week, day, hour, and minute. Important Project monitoring applies to all objects within the specified monitoring scope. New tasks that meet the criteria are automatically configured with the corresponding monitoring job. If a monitored object has both project-level and task-level monitoring configured with the same recipient and notification method, the task-level alert configuration takes precedence. For more information, see Alert rule priorities.
   Alarm Reason	Alert Condition	For project-level monitoring, you can configure alerts for tasks that have an Error, are Not Finished, or have a Timeout. To configure different alert conditions, click + New Alert Condition. Error: An alert is triggered if a task encounters an error during runtime. An alert is triggered if any field in a logical table has an error. Note Dataphin automatically retries a failed task. An alert is triggered only if the task still fails after all retry attempts are exhausted. Not Finished: To configure this alert, set the object type to monitor and the time point, which is the trigger condition. An alert is triggered if the task is not finished by this time. You can monitor Daily/Weekly/Monthly Scheduled Tasks or Hourly/Minutely Scheduled Tasks. You must select at least one type of scheduled task to monitor. Note Changes to an incomplete alert configuration will not take effect until the next day. Daily/Weekly/Monthly Scheduled Tasks: An alert is triggered if the task is not finished within the configured Trigger Condition time. The time zone for the trigger condition is the same as the time zone for the Monitoring Interval parameter. For more information, see the description of the Monitoring Interval parameter below. Hourly/Minutely Scheduled Tasks: The trigger condition can be Not finished after a specific time or Instance not finished after scheduled runtime. Not finished after a specific time: For a specified cycle of a monitored hourly or minutely scheduled task, an alert is triggered if the task is not finished after the specified time. The cycle range is 1 to 288 cycles. To add multiple cycles, click + Add Cycle. You can add a maximum of 24 cycles. The cycles cannot overlap. The time zone for the Monitoring Time Point is the same as the time zone for the Monitoring Interval parameter. For more information, see the description of the Monitoring Interval parameter below. Instance not finished after scheduled runtime: For a specified monitoring cycle of a monitored hourly or minutely scheduled task, an alert is triggered if the instance is not finished after the specified time that follows the scheduled runtime. For the monitoring cycle range, you can select All, First Cycle, Last Cycle, or Custom to select the cycles to monitor. The cycle range is 1 to 288 cycles. Important For Not Finished alerts, changing the scheduling cycle of a monitoring job to hourly or minutely may cause the corresponding monitoring rule to become invalid. An alert is triggered if an unfinished field exists in a logical table at a specified time point. This is valid only for daily tasks. Timeout: To configure this alert, set the timeout duration. The valid range is 0 to 19,999 minutes. If you enter a value outside this range, the system uses the nearest valid value. The timer starts when the instance status changes to Running. An alert is triggered if the task is not finished or if a field in a logical table is not finished after the specified duration.
   	Monitoring Interval	Set the time interval when monitoring is active. Select All Day or Specified Period. When you select Specified Period, you must configure a Start Time and an End Time. If the system time zone, which is the time zone in the User Center, is different from the scheduling time zone, both time zones are displayed. When you select a start time for a specified period, the system automatically calculates the corresponding time in the scheduling time zone. If the calculated time in the scheduling time zone is on a different day than the system time zone, (-1) indicates the previous day and (+1) indicates the next day. For example, if the system time zone is GMT+08:00 and the scheduling time zone is GMT+01:00, and you configure the period as 00:00 ~ 01:00, the corresponding time in the scheduling time zone is 17:00(-1) ~ 18:00(-1).
   	Alert Frequency	The frequency at which alert information is sent. You can set the frequency to a value from 1 to 59 minutes per alert. If you enter a value outside this range, the system uses the nearest valid value.
   	Send Count	The number of times alert information is sent when an alert is triggered. Set an integer from 1 to 10. If you enter a value outside this range, the system uses the nearest valid value.
   	Accept Configuration	The configuration for alert recipients. For the recipient, you can select Owner, Duty Schedule, or Custom. To configure different Recipient settings, click New Recipient. Owner: The owner of the task. Duty Schedule: The on-call schedule configured in the Alert Center. You must create an on-call schedule in advance. For more information, see Manage on-call schedules. Custom: Use this option for scenarios where alerts need to be sent to non-owners and users not on the on-call schedule, such as project administrators. Notification Method: You can select Phone, Text Message, Email, DingTalk, Lark, WeCom, and Custom Channel. Important Select the contact method for the recipient. For more information about how to configure contact methods for recipients, see Add, configure, and manage Dataphin members.

6. Click OK to complete the project monitoring and alert configuration.