Database Autonomy Service (DAS) analyzes your instance's performance data from the last 10 days and predicts CPU utilization for the next 24 hours. When a predicted value is expected to reach the performance bottleneck threshold, DAS generates a scale-out suggestion for you to review.
This feature generates predictions and scale-out suggestions only. Scale-out operations are not performed automatically — you must execute them manually.
Use cases
Prediction-based auto scaling is well suited for:
Cyclical workloads: Databases with predictable load patterns, such as peak traffic during business hours or recurring batch jobs
Proactive capacity planning: Scaling out before a bottleneck occurs, rather than reacting after the fact
Risk-averse scaling: Reviewing forecast data before committing to a specification change, so you scale with confidence
Prerequisites
Before you begin, ensure that you have:
An RDS instance running one of the following versions:
MySQL 8.0 on RDS High-availability Edition, RDS Enterprise Edition, or RDS Cluster Edition, RDS Enterprise Edition,
MySQL 5.7 on RDS High-availability Edition, RDS Enterprise Edition, or RDS Cluster Edition, RDS Enterprise Edition,
MySQL 5.6 on RDS High-availability Edition
MySQL 5.5 on RDS High-availability Edition
Billing
No fees are incurred. Because DAS only predicts and suggests — it never performs scale-out operations automatically — there is no billing impact from enabling this feature.
Set up a prediction-based auto scaling policy
Go to the RDS Instances page. In the top navigation bar, select a region, then click the ID of the target instance.
In the left navigation pane, choose Autonomy Services > Diagnostics > Autonomy Center. On the Autonomy Center page, click Autonomy Service Settings.
On the Autonomous Function Settings > Auto Scaling tab, click Add Policy. Configure the following parameters and click OK.
Parameter Description Policy name A custom name for the policy Mode Select Prediction-based Auto Scaling Engine type Select RDS MySQL Metric type Select CPU Utilization (%) Destination value The CPU utilization threshold that triggers a scale-out suggestion. When the predicted value is greater than or equal to this threshold, DAS generates a suggestion In the Recommended Policies section, find the policy and click Apply in the Actions column. In the dialog box, click Next Step to apply the policy to the instance.
To modify an applied policy, click Modify in its Actions column and update the settings on the Update Policy tab.
To unapply a policy, find it in the Applied Policies section and click Cancel in the Actions column.
On the Autonomous Function Management page, click OK.
(Optional) To receive notifications when a scale-out suggestion is generated, configure alerts and add an alert contact group.
If you have configured an alert template for the instance, follow the prompts to add an alert rule for the corresponding autonomy event to the template.
View prediction results and evaluate suggestions
Go to the RDS Instances page. In the top navigation bar, select a region, then click the ID of the target instance.
In the left navigation pane, choose Autonomy Services > Diagnostics.
On the Autonomy Center page, view Auto-Scaling Events for the selected time range.
Click Details for a Database Workload Prediction(Predict Only) event to see the predicted CPU utilization trend and the scale-out suggestion.
When reviewing a suggestion, consider the following:
How close is the predicted peak to your threshold? A prediction that approaches but does not consistently exceed your threshold may not require immediate action.
Does the pattern repeat? If the forecast shows recurring peaks — for example, every weekday morning — scale out before the next cycle.
How urgent is the risk? The prediction covers the next 24 hours. If a bottleneck is forecast within hours, act promptly.
What's next
Scale out the instance based on the suggestion. For details, see Change instance specifications.