Skip to main content

Creating alerts in Dagster+

You can create alerts in the Dagster+ UI or using the dagster-cloud CLI.

note

Before you create alerts, you must configure an alert notification service.

Alerting when a run fails

You can set up alerts to notify you when a run fails.

By default, these alerts will target all runs in the deployment, but they can be scoped to runs with a specific tag.

  1. In the Dagster UI, click Deployment.

  2. Click the Alert policies tab.

  3. Click Create alert policy.

  4. Select Run from the menu and click Continue.

  5. Choose run targets and events for your alert and click Continue.

  6. Choose a notification channel for your Run alert and click Continue.

  7. Review and save your Run alert and click Save alert.

  8. Select Job failure.

If desired, add tags in the format {key}:{value} to filter the runs that will be considered.

Alerting when a run is taking too long to complete

You can set up alerts to notify you whenever a run takes more than some threshold amount of time.

By default, these alerts will target all runs in the deployment, but they can be scoped to runs with a specific tag.

  1. In the Dagster UI, click Deployment.

  2. Click the Alert policies tab.

  3. Click Create alert policy.

  4. Select Run from the menu and click Continue.

  5. Choose run targets, select "Run exceeding a specified number of hours", enter the number of hours, and click Continue.

  6. Choose a notification channel for your Run alert and click Continue.

  7. Review and save your Run alert and click Save alert.

  8. Select Job running over and how many hours to alert after.

If desired, add tags in the format {key}:{value} to filter the runs that will be considered.

Alerting when an asset fails to materialize

You can set up alerts to notify you when an asset materialization attempt fails.

By default, these alerts will target all assets in the deployment, but they can be scoped to a specific asset or group of assets.

note

If using a RetryPolicy, an alert will only be sent after all retries complete.

  1. In the Dagster UI, click Deployment.
  2. Click the Alert policies tab.
  3. Click Create alert policy.
  4. Select Asset from the menu and click Continue.
  5. Choose asset targets and events and click Continue.
  6. Choose a notification channel for your asset alert and click Continue.
  7. Review and save your asset alert and click Save alert.

Alerting when an asset check fails

You can set up alerts to notify you when an asset check on an asset fails.

By default, these alerts will target all assets in the deployment, but they can be scoped to checks on a specific asset or group of assets.

  1. In the Dagster UI, click Deployment.
  2. Click the Alert policies tab.
  3. Click Create alert policy.
  4. Select Asset from the menu and click Continue.
  5. Choose asset targets, select the relevant asset check events, and click Continue.
  6. Choose a notification channel for your asset alert and click Continue.
  7. Review and save your asset alert and click Save alert.

Alerting when a schedule or sensor tick fails

You can set up alerts to fire when any schedule or sensor tick across your entire deployment fails.

Alerts are sent only when a schedule/sensor transitions from success to failure, so only the initial failure will trigger the alert.

  1. In the Dagster UI, click Deployment.
  2. Click the Alert policies tab.
  3. Click Create alert policy.
  4. Select Automation from the menu and click Continue.
  5. Select the tick types to be alerted on. Choose if you only want to be alerted after n consecutive failures and click Continue.
  6. Choose a notification channel for your automation alert and click Continue.
  7. Review and save your automation alert and click Save alert.

Alerting when a code location fails to load

You can set up alerts to fire when any code location fails to load due to an error.

  1. In the Dagster UI, click Deployment.
  2. Click the Alert policies tab.
  3. Click Create alert policy.
  4. Select Code Location from the menu and click Continue.
  5. Select your target and click Continue.
  6. Choose a notification channel for your code location alert and click Continue.
  7. Review and save your code location alert and click Save alert.

Alerting when a Hybrid agent becomes unavailable

note

This is only available for Hybrid deployments.

You can set up alerts to fire if your Hybrid agent hasn't sent a heartbeat in the last 5 minutes.

  1. In the Dagster UI, click Deployment.
  2. Click the Alert policies tab.
  3. Click Create alert policy.
  4. Select Code Location from the menu and click Continue.
  5. Select your target and click Continue.
  6. Choose a notification channel for your code location alert and click Continue.
  7. Review and save your code location alert and click Save alert.