Function Compute:Configure provisioned instances

Step 1: Create a provisioned instance policy

1. Log on to the Function Compute console. In the left-side navigation pane, click Function.

2. In the top navigation bar, select a region. On the Function page, click the name of the function that you want to manage.

3. On the Function Details tab, click the Configuration tab.

4. In the left-side navigation tree, click the Provisioned Instances tab and click Create Provisioned Instance Policy.

5. In the Create Provisioned Instance Policy panel, configure the parameters and click OK.

   Parameter	Description
   Version or Alias	Select the version or alias for which you want to create a provisioned instance policy. Note You can create a provisioned instance policy only for the LATEST version.
   Provisioned Instances	Specify the number of provisioned instances. Note The minimum number of provisioned instances help quickly respond to function invocation requests, reduce cold starts, and improve service performance for online applications that are sensitive to response latency. Take note that you are charged for these instances even if they do not process any request unless you release them.
   (Optional) Scheduled Scaling: You can select this option to configure a scheduled scaling policy which scales the number of function instances at a specified point in time. For more information about the scenario and configuration example, see Scheduled scaling.
   Policy Name	Enter a custom policy name.
   Provisioned Instances	Specify the number of instances that you want to scale out. Note After you configure this parameter, its value overrides the Provisioned Instances value configured earlier in this section.
   Trigger Mode	You can select At Time Points or Custom CRON expression. At Time Points: Specify the Time (UTC), Date (UTC), and Week Day (UTC) parameters as prompted. Custom CRON expression: Specify the Schedule Expression (UTC) parameter. In this example, cron(0 0 4 * * *) is configured to trigger scaling at 12:00 (UTC+8) every day. Note Because the time setting must be in UTC time, the time is set to 4:00 in this example.
   Effective Time (UTC)	Set the time when the scaling configurations start to take effect and the time when the scaling configurations expire.
   (Optional) Water-level Scaling: You can select this option to scale function instances every minute based on the concurrency usage metric or another usage metric. For more information about the scenario and configuration example of water-level scaling, see Water-level scaling.
   Policy Name	Enter a custom policy name.
   Minimum Number of Provisioned Instances	Specify the minimum number of instances that you want to scale.
   Maximum Number of Provisioned Instances	Specify the maximum number of instances that you want to scale.
   Utilization Type	Note This parameter is displayed only when GPU-accelerated instances are configured. Select a metric based on which instances are scaled.
   Concurrency Usage Threshold/Usage Threshold	Configure a usage threshold. If the usage of a metric for an instance or the concurrent usage falls below the configured threshold, a scale-in is triggered. If the usage of a metric for an instance or the concurrent usage reaches or exceeds the configured threshold, a scale-out is triggered.
   Effective Time (UTC)	Set the time when the scaling configurations start to take effect and the time when the scaling configurations expire.

   After the policy is created, you can view the policy in the policy list of the function.

Step 2: Verify the policy

You can check whether the configured policy takes effect by checking the number of provisioned instances in the monitoring data when specific condition is fulfilled.

1. On the Function Details tab, click the Monitoring tab.

2. On the Function Metrics tab, view the data in the Function Provisioned Instances (count) card to check whether the policy takes effect.