Create a scenario
Procedure
1. Log on to the Personalized Algorithm Development Platform console.
2. In the navigation pane on the left, click Personalized Algorithm Development Platform > Instance List.
3. In the top menu bar, select the region where the instance is located.
4. Click the instance ID, click the Scenario Management menu, and then click Create Scenario. The following table describes the parameters for creating a scenario.
Parameter | Description |
Clone from an existing scenario | Lets you quickly clone an existing scenario within an instance. The basic information of the scenario is also cloned. If you are creating a scenario for the first time, do not select this option. |
Scenario name | The display name of the scenario. |
Scenario response time (RT) limit | The unit is ms. After you set the RT limit, all requests in the scenario that exceed this time fail. Enter this value with care. Do not set the value to more than 999 ms. |
Diversion rule | Use the parameters passed to TPP in your request as the hash factor for scenario experiment diversion. If the hash factor is not passed or is empty, the request is randomly assigned to a bucket. Hash factor recommendations: If a high percentage of your users are logged on, use the user ID for diversion. If a high percentage of users are not logged on, use the device ID for diversion. |
5. After you enter the basic information for the scenario, click OK to create the scenario.
Note: Creating a scenario involves allocating resources, such as CPU and memory. Each scenario is isolated and requires a separate resource request. If the required resources exceed the CPU and memory that you purchased, the allocation fails, and the scenario is not created. After a scenario is created, its status is Running. If the creation fails, its status is Not Published.
Possible error | Meaning | Solution |
Resource allocation log: java.lang.IllegalStateException: idc cn-shanghai-a is not exist in cm2 idc configs | Resource allocation failed. | Check if resources are sufficient. Check if the region where the instance is located has enough resources. If scenario resources are insufficient, scale out the instance to add more resources. |