Skip to main content
An alert brings you the information you need to triage, prioritize and debug an issue. Alerts use your ownership, data product and monitor settings. You can route them to several channels by rules such as data warehouse schemas, metadata tags or specific tables. Coalesce Quality sends alerts for monitors (anomaly, freshness and volume), transformation layers (dbt, SQLMesh and Coalesce Transform test failures, model run failures and source freshness issues), Airflow task and DAG failures, and schema changes. Alerts go to Slack, email, Microsoft Teams and webhooks, including PagerDuty and Opsgenie. You can send an alert straight to the channels you choose, or attach it to owners so it follows your ownership model. See Routing alerts. Example anomaly monitor alert title

Core concepts

  • New and ongoing issues. By default you are alerted only about new issues. You can also get reminders about issues that are still open. See Configuring an alert.
  • Ownership. If the asset has an owner, the alert shows it. If the owner has a Slack handle, the person or group is tagged in the alert.
  • Recent code changes. The alert shows the most recent code change on the asset and who made it, so you can check if it caused the error. Click it to open the change in GitHub or GitLab.
  • Downstream assets. The alert shows how many downstream tables and BI assets depend on the asset, to help you judge the impact. Click it to open a filtered lineage view.
  • Error message. The alert shows the error message from the source system, cleaned up where possible. For dbt and Coalesce Transform tests, it shows what is wrong and how many rows failed.
  • Anomaly chart. For an anomaly, the alert shows the chart with past annotations.
  • Status. You can change the issue status from the alert, so others can see who is working on it.
A Slack message shows at most 15 issues per section. If there are more, the message ends the section with “… and X more”.

Setting up an alert

Routing alerts

Configuring an alert