Scheduled auto scaling

更新时间:
复制 MD 格式

Database Autonomy Service (DAS) provides scheduled auto scaling policies for Tair (Redis OSS-compatible). These policies scale out your database instance according to a preset schedule and automatically revert the instance to its original specifications after the configured duration ends. Use this feature to handle predictable, periodic changes in database load while meeting business needs and controlling costs.

Prerequisites

  • The instance runs Redis Open-Source Edition or memory-optimized Tair (Enterprise Edition).

  • The instance uses the standard architecture.

  • The instance is a cloud-native edition.

  • You have created a DAS service-linked role.

  • Your account balance must be sufficient to cover the cost of scaling out.

Important notes

  • This feature currently supports only scheduled scale-out operations based on preset policies.

  • All time settings use UTC+8. If your database instance is in a different time zone, convert the times accordingly.

  • Only one scaling policy of the same type can be applied to each instance.

  • Scheduled auto scaling does not support automatic retry. If a one-time scaling rule fails, it will not be retried.

  • If you modify the Duration or the Scale-back Time after scaling out, the instance reverts to its original specifications based on the latest configuration.

  • Setting a Duration or Scale-back Time does not guarantee that the instance reverts to its pre-policy specifications:

    • If another specification change occurs during this period—such as a manual adjustment or dynamic scaling—and the instance’s current specifications no longer match the target specifications defined in the scheduled policy, the instance will not revert.

    • To ensure business stability and safety, the instance reverts only when relevant metrics meet downgrade conditions. For example, if the original instance has 1 GB of memory and the scheduled policy scales it to 4 GB, but the current memory usage is 1 GB at restore time, the resulting memory usage would be 1 GB / 1 GB = 100%. Since this exceeds safe thresholds, the instance will not revert.

    • If the instance is already undergoing a specification change, migration, or any other operation that prevents modification, the restore operation cannot execute.

Use cases

DAS can scale up an instance once or on a recurring basis, such as daily, weekly, or monthly. For example, a business experiences a peak period from 2:00 on the 1st to 2:00 on the 3rd of every month. You can use the scheduled auto scaling feature to automatically scale up the instance to a higher specification at 2:00 on the 1st and then automatically restore it to the original specification after 2:00 on the 3rd.

Billing information

When scheduled auto scaling changes your database instance specifications, the target instance incurs corresponding upgrade/downgrade fees. For details, see Upgrade/Downgrade.

Create a scheduled auto scaling policy

  1. Log on to the console and go to the Instances page. In the top navigation bar, select the region in which the instance that you want to manage resides. Then, find the instance and click the instance ID.

  2. In the navigation pane on the left, click CloudDBA > Performance Trends.

  3. On the Performance Trends page, click Autonomy Service Settings.

  4. On the Autonomous Function Settings > Auto Scaling tab, click Add Policy. On the Add Policy page, configure the following parameters:

    Parameter

    Description

    Policy Name

    The name of the policy.

    Mode

    Select Scheduled Auto Scaling.

    Engine Type

    The database engine type.

    Specifications

    The specifications of the selected database engine.

    Operation

    Select Adjust Instance Specifications to the target size.

    Note

    Currently, only Adjust Instance Specifications is supported.

    Valid From

    Valid From includes Start date and End date.

    • Start date is required and must be today or later.

    • End date is optional.

      • If Recurrence is set to No recurrence (one-time execution), the policy end time is not constrained by the End date.

      • If Recurrence is set to Daily, Weekly, or Monthly, and you do not set an End date, the policy runs indefinitely. If you set an End date, the policy stops after that date.

    Recurrence

    Options:

    • No recurrence (one-time execution)

      • Scaling Start Time: Required.

      • Duration: Optional. Enter a positive integer in hours.

        • If you do not set Duration, the instance does not revert to its original specifications after the policy executes.

        • If you set Duration, the instance reverts to its original specifications after the duration ends.

    • Daily:

      • Scaling Start Time: Required.

      • Scale-back Time: Required.

        • If Scaling Start Time < Scale-back Time, the policy starts and ends on the same day.

        • If Scaling Start Time > Scale-back Time, the policy starts on one day and ends the next. The Scaling Start Time applies to the current day, and the Scale-back Time applies to the following day.

        Note
        • Scaling Start Time and Scale-back Time must be at least 1 hour apart.

        • Scaling Start Time and the Scale-back Time of the previous epoch's scheduled auto scaling policy must be at least 1 hour apart.

        • If an effective end date is set, and the effective end date falls between the Scaling Start Time and the Scale-back Time of a cycle, the scheduled auto scaling policy for that cycle will not be executed.

    • Weekly

      • Scaling Start Time: Required.

      • Scale-back Time: Required.

        • If Scaling Start Time < Scale-back Time, the policy starts and ends within the same week.

        • If Scaling Start Time > Scale-back Time, the policy starts in one week and ends the next. The Scaling Start Time applies to the current week, and the Scale-back Time applies to the following week.

        Note
        • Scaling Start Time and Scale-back Time must be at least 1 hour apart.

        • Scaling Start Time and the Scale-back Time of the previous epoch's scheduled auto scaling policy must be at least 1 hour apart.

        • If you set an Effective End Date, and the Effective End Date falls between the Scaling Start Time and Scale-back Time of a cycle, the scheduled auto scaling policy for that cycle will not execute.

    • Monthly

      • Adjust start time (monthly): Required.

      • Restore time (monthly): Required.

        • If Scaling Start Time < Scale-back Time, the policy starts and ends within the same month.

        • If Scaling Start Time > Scale-back Time, the policy starts in one month and ends the next. The Scaling Start Time applies to the current month, and the Scale-back Time applies to the following month.

        Note
        • Scaling Start Time and Scale-back Time must be at least 1 hour apart.

        • Scaling Start Time and the Scale-back Time of the previous epoch's scheduled auto scaling policy must be at least 1 hour apart.

        • If you set an Effective End Date, and the Effective End Date falls between the Scaling Start Time and the Scale-back Time of a cycle, the scheduled auto scaling policy for that cycle will not be executed.

  5. In the Recommended Policies section, click Apply in the Actions column to add the policy to the instance.

    Note
    • To modify an existing policy, click Modify in the Actions column and update the policy on the Update Policy tab.

    • To remove an applied policy, go to the Applied Policies section and click Cancel in the Actions column.

  6. Click OK.

View scheduled auto scaling results

  1. Log on to the DAS console.

  2. In the navigation pane on the left, click Instance Monitoring.

  3. Find the target instance and click its instance ID to open the instance details page.

  4. In the navigation pane on the left, click Autonomy Center.

  5. On the Autonomy Center page, view Auto-Scaling Events that occurred during your selected time range.

  6. Click Details for an Auto-Scaling Events to view detailed information about the automatic scaling activity.

FAQ

Q: What should I do if my instance has reached the maximum specification for its series and cannot scale out?

A: Purchase a higher-spec instance from another series and migrate your data to the new instance. For instance specifications and migration guidance, see Instance specifications and Instance migration.