Traffic layers
A traffic layer is a container that holds and allocates experiment traffic. It is mainly used to:
- Make experiments mutually exclusive and isolate traffic.
- Ensure that every experiment is bound to a traffic layer.
- Ensure that a user matches at most one experiment rule within the same traffic layer.
- Automatically release the traffic occupied by an experiment after the experiment ends.
Create a traffic layer
Click Create Traffic Layer in the upper-right corner and fill in the traffic layer information. You can expand the feature description at the top of the pop-up.
| Field | Description |
|---|---|
| Name | Business name used to identify the traffic layer |
| Group | Used for categorized management; defaults to Default |
| Split-flow Mode | Currently only Remote Split-flow is supported. Local Split-flow is not available yet |
| Bucket Key | Determines which identifier is used to group users consistently |
| Remarks | Records the usage scope and mutual exclusion purpose of the traffic layer |
After you click Save, the system generates a traffic layer ID.
View traffic layer details
Click a traffic layer name to view Basic Information, Associated Experiments, and Finished Experiments.
After an experiment ends, it moves to Finished Experiments in the traffic layer details, and the traffic it occupied is released automatically.
If you plan to run multiple experiments in parallel in the same traffic layer, reserve available traffic for later experiments. If two experiments must not exclude each other, put them in different traffic layers.