Alerts defined

The Alerts defined page lists all alert definitions that have been created within your Databand environment. Alert definitions can be filtered using the Add filters button at the top left of the page. The filter dimensions below are available:

  • Assets: The sources, projects, or pipelines in your environment.
  • Alert type: The specific types of alerts that are available through Databand.
  • Severity: A user-defined severity attribute that can be attached to any alert definition.
  • Collaborators: The individuals who have been identified as collaborators for the alert.

Alert metadata

The list of alerts provides various pieces of metadata about each alert. In general, all alerts provide the following information at a minimum:

  • Alert name, type, severity
  • Any receivers or collaborators that are attached to the alert
  • Assets associated with the alert (for example, source, project, pipeline, or dataset)
  • The time when the alert was created or modified

Alert types

In Databand, alerts are separated into three distinct types: pipeline, task, and dataset alerts. The following sections include all of the possible alert definitions that you can set. To create alerts, see managing alerts.

Pipeline alerts

For pipelines, you can create the following alert definitions:

  • Pipeline state: Know when your pipeline has failed, succeeded, or entered some other specific state.
  • Pipeline duration: Identify pipeline durations that fall outside of defined limits or are anomalous.
  • dbt test: Know when any of your pipeline's dbt tests have failed.
  • Schema change: Compare dataset schemas between runs in order to identify any cases where columns have been added or removed or column data types have changed.
  • Pipeline SLA: Know when your pipeline has not started or completed on time according to SLAs.

Task alerts

For tasks, you can create the following alert definitions:

  • Task state: Know when your task has failed, succeeded, or entered some other specific state.
  • Task duration: Identify task durations that fall outside of defined limits or are anomalous.
  • Custom task metric: Create an alert for any metric captured by Databand. This alert type serves as a catch-all for cases where a streamlined alerting process may not be available for a particular type of metric.

Dataset alerts

For datasets, you can create the following alert definitions:

  • Missing operation: Identify dataset operations that are missing from a pipeline run compared to historical runs.
  • Tables data quality: Set up data quality checks for the tables monitored in your database.
  • Operations data quality: Set up data quality checks for the datasets logged in your pipelines.
  • Data delay: Establish daily or hourly targets for the completion of dataset updates.
  • Data quality metric: Set up data quality checks for SQL queries.