Data Observability alerts
With Data Observability alerts, you can track all metadata from job states to data quality metrics for DataStage jobs. You can track the job execution and receive information about issues with your data as soon as they occur.
By using the broad alerting and analysis capabilities that Data Observability provides, you can observe and analyze the progress of all running workflow instances that you are observing. As a result, you can make sure that your data team is notified about job condition and the quality of inputs and outputs.
Data Observability roles and permissions
You can display alerts and jobs for your workspaces only. To work with alerts, you need the following permissions and roles:
| Permission | Tasks |
|---|---|
| Manage alert | Create, edit, and delete alerts for assets from the workspaces to which you have access. You can manage alerts only in the workspaces in which you have an Admin or Editor role. |
| View triggered alerts | View all triggered alerts and the related details. |
| View defined alerts | View defined alerts and the related details. |
| Acknowledge and resolve alerts | Indicate that you are aware of the alert and resolve the triggered alerts. |
The following permissions are granted by default when you have the access to the workspace.
- View triggered alerts
- View defined alerts
- Acknowledge and resolve alerts
The Manage alert permission is granted automatically when you are a project Admin or Editor (in the project you have the assets that you want to observe).
Contact the Workspace Admin to apply for the Manage alert permission, Admin role, or Editor role.
Alert definition types
You can create different types of alerts. Look at the following list and choose the alert type that best fits your requirements:
-
Threshold metric alerts: Select this type of an alert to observe the duration of your jobs (duration alerts) or any other specific parameter value (metric alerts).
- A Job run duration alert is triggered when the duration of job runs doesn't meet the defined time or is anomalous.
- A System metric alert is triggered when there are differences in metrics between the source and target datasets. These differences are identified at the individual metric level. These metrics can include:
- Average throughput
- Rows read
- Rows written
- Elapsed time
-
State alerts: Use this alert if you need to monitor the state of the completed job run. A Job run state alert is triggered when a job run is completed with the state that is defined in the alert.