Traffic spikes can push Tair (Redis OSS-Compatible) instance memory to its limit before you have time to act. The automatic scale-up feature in Database Autonomy Service (DAS) monitors average memory usage and scales up the instance when usage reaches a threshold you set—preventing memory overflows without requiring manual intervention.
Limitations
Before you enable this feature, confirm that your instance meets the following requirements:
-
Deployed in cloud-native mode
-
Using the standard architecture. For more information, see Standard architecture
Billing
You are charged only for each scale-up (configuration change). For more information, see Configuration changes.
How it works
After you enable the feature, DAS monitors the average memory usage of the instance within the observation window you configure. When usage reaches the threshold, DAS doubles the memory size—for example, from 1 GB to 2 GB. DAS then continues monitoring. If the threshold is reached again, DAS performs another scale-up. This continues until the instance reaches the maximum memory size you specified.
DAS does not scale down automatically. When average memory usage drops below 30% within the observation window, DAS sends a scale-down suggestion through your configured notification channel (such as email). Scale down the instance manually at a time that suits your workload to reclaim costs.
-
When you enable this feature, the system authorizes DAS to assume the AliyunServiceRoleForDAS role to access your instance resources.
-
If you make a manual configuration change after an automatic scale-up, DAS treats it as a manual scale-down and stops the scale-down suggestion cycle. DAS resumes sending scale-down suggestions only after the next automatic scale-up triggers and average memory usage drops below 30% again.
Impacts
Expect the following during a scale-up:
| Impact | Details |
|---|---|
| Transient connections | The instance may experience one or two transient connections, each lasting up to 30 seconds. |
| Read-only period | The instance remains read-only for up to 1 minute during the configuration change. This syncs incremental data and prevents dual writes caused by Domain Name System (DNS) caching, ensuring data consistency between the original and new instance. |
| Minor version update | If the instance is running an outdated minor version, the system updates it to the latest minor version during the scale-up. |
Enable automatic scale-up
-
Log on to the DAS console.
-
In the left-side navigation pane, choose Intelligent O&M Center > Instance Monitoring.
-
Find the instance and click its instance ID to open the instance details page.
-
In the left-side navigation pane, click Autonomy Center.
-
In the upper-right corner of the Autonomy Center tab, click Autonomy Service Settings.
-
On the Autonomous Function Settings tab, click Optimization and Throttling, select Automatic Scale-up/out, and configure the following parameters. Example: If average memory usage is >= 70% within a 30-minute observation window, DAS doubles the instance memory size. The memory can increase to up to 64 GB.
Parameter Description Average memory usage The memory usage threshold that triggers a scale-up. Valid values: 50–90 (%). Adjust at a step of 10%. Maximum specifications The maximum memory size to scale up to. DAS stops scaling after the instance reaches this size. Observation window The monitoring window length. Maximum: 30 minutes. -
(Optional) Click the Event Subscription Settings tab to configure event notifications so DAS can alert you when scale-up events or scale-down suggestions occur. Urgency levels:
Parameter Description Enable subscription service Turn on to receive event notifications. Urgency The minimum event level that triggers a notification. To receive automatic scale-up events and scale-down suggestions, select Notification. For all available levels, see the table below. Select contacts/contact groups Select contacts or contact groups from the Existing Contacts/Contact Groups section and click the arrow icon to move them to the Selected Contacts/Contact Groups section. If no contacts exist, create one first. For more information, see Manage alert contacts. Notification method The notification channel. Select one or more methods. Include SMS for time-sensitive delivery. Notification language The language for notifications. Only Chinese is supported. Minimum notification interval The minimum time between two notifications about the same event. For example, setting this to 5 minutes means you receive at most one notification per 5-minute window for the same event. Level Events included Critical Events that affect database operation Warning Events that may affect database operation Optimize Optimization suggestions based on database status Notification Basic events and performance snapshot events