View a markdown version of this page

Step 3: Starting an experiment - Amazon AppConfig
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Step 3: Starting an experiment

Before you start an experiment, verify you completed the prerequisites. For more information, see Step 1: Configuring prerequisites. Before running an experiment in production, read About running and monitoring an experiment to familiarize yourself with recommended practices for safely starting and monitoring the experiment.

To start an experiment

  1. Open the Amazon Systems Manager console at https://console.amazonaws.cn/systems-manager/appconfig/.

  2. In the navigation pane, choose Experiments, and then choose an experiment to run. The experiment definition page opens.

  3. Choose Start experiment run.

  4. (Optional) In the Description field, enter information about this particular experiment run.

  5. For an initial experiment run, verify that Test first at 0% exposure on start is selected.

  6. If you plan to use treatment assignment overrides, choose Add entity ID. For more information, see About treatment assignment overrides.

  7. (Optional) Expand Tags to assign one or more key-value pairs to the experiment run.

  8. (Optional) Expand Deployment parameters to assign a Amazon Key Management Service (Amazon KMS) key and tags to the experiment run.

  9. For an initial experiment run, leave audience exposure set to 0%.

  10. Choose Start experiment run. Amazon AppConfig experimentation opens the dashboard for the running experiment.

At 0% exposure, no audience traffic is assigned to treatments unless overrides are configured. Use this stage to validate feature flags, treatment logic, metrics, and logging before increasing exposure. For more information, see Operational considerations.