Feature management
View the Feature list
Go to Config Center → Feature. The list shows each Feature's parameter name, display name, status, parameter type, default value, assignment rules, group, integration side, description, creator, last modifier, and actions.
Feature statuses include:
- Draft: The Feature has been created but doesn't participate in rule evaluation yet.
- Enabled: The Feature participates in rule evaluation and can be used to create Feature experiments.
Create a Feature
Click Create Feature in the upper-right corner and fill in the Feature information.
| Field | Description |
|---|---|
| Feature parameter name | Identifies the Feature. Use a stable parameter name that reflects its business meaning |
| Display Name | A name that helps business users recognize the Feature |
| Parameter type | String, Number, Boolean, or JSON |
| Default value | The fallback value used when a user doesn't match any custom rule |
| Integration side | Client or Server-side |
| Group | Used for categorized management. If no group is selected, Default is used |
| Remarks | Records the business purpose of the Feature |
When you're done, click Create. A new Feature starts in the Draft status.
Click the run icon in the Feature's action column, then click Enable in the confirmation window. After it's enabled, the Feature status changes to Enabled and the Feature starts participating in rule evaluation.
Create experiment rules and edit the default value
For an enabled Feature, the + menu in the action column provides three shortcuts: Whitelist Rule, Experiment Rule, and Edit Default Value.
- Experiment Rule: Opens the Feature experiment editing page with the current Feature preselected. You still need to enter the experiment name, select a traffic layer, set the traffic between groups and the Feature values, and add metrics before you submit the experiment.
- Edit Default Value: Adjusts the fallback value of the Feature. The default value applies to users who don't match any custom rule. You can clear the default value. After it's cleared, online users who get this Feature through the default rule no longer receive it in their pull results, so proceed with caution.
Changing the default value affects users who don't match other rules. Save the change only after you confirm that the client is compatible with the new value. When an experiment ends, the Feature assignment falls back to the rule with the next priority. If there are no other rules, the default value is used.
The Feature list also provides an archive entry. Before you archive a Feature, make sure it's no longer used by running experiments or business rules. You can view archived content through Archived in the upper-left corner of the page.