See the current evaluation rules and audience traffic for a feature - Amazon CloudWatch

See the current evaluation rules and audience traffic for a feature

Important

End of support notice: On October 16, 2025, AWS will discontinue support for CloudWatch Evidently. After October 16, 2025, you will no longer be able to access the Evidently console or Evidently resources.

You can use the CloudWatch Evidently console to see how the feature's evaluation rules are allocating the audience traffic among the feature's current launches, experiments, and variations.

To view the audience traffic for a feature

  1. Open the CloudWatch console at https://console.aws.amazon.com/cloudwatch/.

  2. In the navigation pane, choose Application Signals, Evidently.

  3. Choose the name of the project that contains the feature.

  4. Choose the Features tab.

  5. Choose the name of the feature.

    In the Evaluation rules tab, you can see the flow of audience traffic for your feature, as follows:

    • First, the overrides are evaluated. These specify that certain users are always served a specific variation. The sessions of users who are assigned overrides do not contribute to launch or experiment metrics.

    • Next, the remaining traffic is available for the ongoing launch, if there is one. If there is a launch in progress, the table in the Launches section displays the launch name and the launch traffic split among the feature variations. On the right side of the Launches section, a Traffic indicator displays how much of the available audience (after overrides) is allocated to this launch. The rest of the traffic not allocated to the launch flows to the experiment (if any) and then the default variation.

    • Next, the remaining traffic is available for the ongoing experiment, if there is one. If there is an experiment in progress, the table in the Experiments section displays the experiment name and progress. On the right side of the Experiments section, a Traffic indicator displays how much of the available audience (after overrides and launches) is allocated to this experiment. The rest of the traffic not allocated to the launch or the experiment is served the default variation of the feature.