Configure intelligent tuning for Azure Database for PostgreSQL - Flexible Server by using the Azure CLI

APPLIES TO: Azure Database for PostgreSQL - Flexible Server

You can verify and update the intelligent tuning configuration for an Azure Database for PostgreSQL flexible server instance by using the Azure CLI.

To learn more about intelligent tuning, see the overview.

Prerequisites

  • If you don't have an Azure subscription, create a trial account before you begin.

  • Install or upgrade the Azure CLI to the latest version. See Install the Azure CLI.

  • Sign in to your Azure account by using the az login command. Note the id property, which refers to the subscription ID for your Azure account.

    az login
    
  • If you have multiple subscriptions, choose the appropriate subscription in which you want to create the server by using the az account set command:

    az account set --subscription <subscription id>
    
  • If you haven't already created an Azure Database for PostgreSQL flexible server instance, create one by using the az postgres flexible-server create command:

    az postgres flexible-server create --resource-group myresourcegroup --name myservername
    

Verify current settings

Use the az postgres flexible-server parameter show command to confirm the current settings of the intelligent tuning feature.

You can verify if this feature is activated for the server mydemoserver.postgres.database.chinacloudapi.cn under the resource group myresourcegroup by using the following command:

az postgres flexible-server parameter show --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning --query value

You can inspect the current setting of the intelligent_tuning.metric_targets server parameter by using the following command:

az postgres flexible-server parameter show --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning.metric_targets --query value

Enable intelligent tuning

To enable or disable intelligent tuning, use the az postgres flexible-server parameter set command. You can choose among the following tuning targets: none, Storage-checkpoint_completion_target, Storage-min_wal_size,Storage-max_wal_size, Storage-bgwriter_delay, tuning-autovacuum, and all.

Important

Autovacuum tuning is currently supported for the General Purpose and Memory Optimized server compute tiers that have four or more vCores. The Burstable server compute tier isn't supported.

  1. Activate the intelligent tuning feature by using the following command:

    az postgres flexible-server parameter set --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning --value ON
    
  2. Select the tuning targets that you want to activate.

    • To activate all tuning targets, use the following command:

      az postgres flexible-server parameter set --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning.metric_targets --value all
      
    • To enable autovacuum tuning only, use the following command:

      az postgres flexible-server parameter set --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning.metric_targets --value tuning-autovacuum
      
    • To activate two tuning targets, use the following command:

      az postgres flexible-server parameter set --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning.metric_targets --value tuning-autovacuum,Storage-bgwriter_delay
      

    If you want to reset a parameter's value to the default, simply exclude the optional --value parameter. The service then applies the default value. In the preceding example, the command would look like the following and would set intelligent_tuning.metric_targets to none:

    az postgres flexible-server parameter set --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning.metric_targets
    

Note

Both intelligent_tuning and intelligent_tuning.metric_targets server parameters are dynamic, meaning no server restart is required when their values are changed.

Considerations for selecting values for tuning targets

When you're choosing values from the intelligent_tuning.metric_targets server parameter, take the following considerations into account:

  • The NONE value takes precedence over all other values. If you choose NONE alongside any combination of other values, the parameter is perceived as set to NONE. This is equivalent to intelligent_tuning = OFF, so no tuning occurs.

  • The ALL value takes precedence over all other values, with the exception of NONE. If you choose ALL with any combination, barring NONE, all the listed parameters undergo tuning.

  • The ALL value encompasses all existing metric targets. This value also automatically applies to any new metric targets that you might add in the future. This allows for comprehensive and future-proof tuning of your Azure Database for PostgreSQL flexible server instance.

  • If you want to include another tuning target, you need to specify both the existing and new tuning targets. For example, if bgwriter_delay is already enabled and you want to add autovacuum tuning, your command should look like this:

    az postgres flexible-server parameter set --resource-group myresourcegroup --server-name mydemoserver --name intelligent_tuning.metric_targets --value tuning-autovacuum,Storage-bgwriter_delay
    

    Specifying only a new value would overwrite the current settings. When you're adding a new tuning target, always ensure that you include the existing tuning targets in your command.

Next steps