Skip to main content

Custom metrics

Dagster+ feature

This feature is only available in Dagster+.

Early access preview

This feature is in preview phase and is available in Dagster+ in limited early access. Functionality and APIs may change as we continue development. To get early access to this feature, reach out to your Dagster account team. For more information, see the API lifecycle stages documentation.

Custom metrics let you track any numeric asset metadata in Dagster+ Insights. Unlike built-in metrics such as run duration and failure counts, custom metrics are defined by you and powered by the numeric metadata your assets emit at runtime. Metric values are ingested in near real-time and displayed as time series alongside built-in metrics in the Insights UI.

Prerequisites​

  • A Dagster+ account
  • The Editor role or above in your deployment

Step 1: Emit numeric metadata on your assets​

Custom metrics are backed by asset metadata. You'll need one or more assets that emit the same metadata key with numeric (int or float) values at materialization or observation time.

For example, the following asset emits a processing_time_ms metric each time it materializes:

import dagster as dg

@dg.asset
def my_table():
    elapsed_ms = 1500  # your logic here
    return dg.MaterializeResult(
        metadata={"processing_time_ms": elapsed_ms}
    )

For more details, see the metadata guide.

Step 2: Create a custom metric​

  1. In the Dagster UI, navigate to Deployment > Custom metrics.
  2. Click Add custom metric.
  3. Fill in the following fields:
    • Metadata key (required) — must exactly match the metadata key your assets emit. This field cannot be changed after creation.
    • Display name (optional) — a human-readable name shown in the Insights UI.
    • Description (optional) — additional context about what the metric tracks.
note
  • Each deployment supports up to 500 custom metrics.
  • Non-numeric metadata values for a registered key do not count against this limit.

Step 3: View custom metrics in Insights UI​

Once created, custom metrics appear alongside built-in metrics in the Insights UI. In the Insights sidebar navigation, custom metrics are grouped under the User provided metrics category.

Monitoring custom metrics with alerts​

You can create alert policies to monitor your custom metrics:

  1. In Dagster, navigate to Deployment > Alert policies and click Create alert policy.
  2. Select Asset as the alert type.
  3. Choose Metrics as the event type.
  4. Select your custom metric from the metric dropdown.
  5. Configure thresholds (greater than or less than a value).

For the full list of alert configuration options, see Creating alert policies.

Managing custom metrics​

You can edit a custom metric's display name and description at any time from the Deployment > Custom metrics page in Dagster. The metadata key cannot be changed after creation — to track a different key, create a new metric.

To delete a custom metric, click the menu on the metric row and select Delete. Deleting a metric removes it from Insights views but does not affect the underlying asset metadata.

Permissions

Those with Editor role and above can create, edit, and delete custom metrics. Viewer and Launcher roles can view custom metrics in Insights, but cannot modify them.