Scaling strength measures the health of your resource pools and predicts scale-out success. By adding multiple zones and alternative instance types, you can reduce scale-out failures and maintain service stability.
Key concepts
Elastic strength measures the scaling capability of an ECS-type scaling group by evaluating the inventory health score of each resource pool.
-
Resource pool
A resource pool is a combination of specified zones and instance types. During a scale-out, the scaling group creates instances from these pools based on predefined rules. The diversity and health of your resource pools directly determine the scale-out success rate.
ImportantNot all resource pools are available during scale-out events
During a scale-out, the scaling group filters resource pools based on scaling configuration constraints (for example, some instance types do not support ESSDs) and creates instances only from the filtered pools.
-
Inventory health score
The inventory health score reflects the supply stability of a single resource pool. It combines three dimensions: current inventory sufficiency, future replenishment capability, and resource popularity. A higher score indicates more reliable supply.
-
5–6 points (High certainty): Resource supply is sufficient, and scale-out success rate is high.
-
1–4 points (Uncertain supply): Resource supply is uncertain. Consider adding vSwitches in other zones or alternative instance types to your resource pool.
-
-3–0 points (Supply warning): Resource supply is low, with a high risk of scale-out failure. Adjust your configuration promptly.
-
View scaling strength
View and adjust scaling strength for a scaling group.
Log on to the Auto Scaling console.
In the navigation pane on the left, click Scaling Groups.
In the top navigation bar, select a region.
-
Find the scaling group and check its scaling strength.
The scaling strength appears next to each scaling group in the list.
-
Click the scaling strength icon to view the inventory health score for each combination of Instance Type and vSwitch. Click a specific score to view calculation details.
In the scaling strength details dialog box, a table displays columns for Instance Type, Zone, vSwitch List, Status, and Inventory Health. A tip at the top suggests improving the scaling strength by selecting vSwitches in different zones for the scaling group.
Best practices
-
Improve scaling strength with resource diversity
Avoid relying on a single resource pool. Greater resource diversity helps scaling groups find alternatives when a specific resource is unavailable. Expand your fallback options in the following ways:
-
Configure multiple zones: When creating a scaling group, associate vSwitches from at least two zones. If resources are constrained in one zone, Auto Scaling automatically attempts to create an instance in another zone.
-
Add alternative instance types: Add multiple alternative instance types to your scaling configuration instead of configuring only one. This allows the system to use a different type when one has insufficient inventory.
-
-
Continuously monitor and optimize scaling strength
Scaling strength is a dynamic metric. A one-time "perfect" configuration does not guarantee long-term effectiveness.
For example, scaling strength can decrease due to zone decommissioning, instance type retirement, or changes in inventory capacity.
Incorporate scaling strength evaluation into your regular O&M procedures. Before major events or promotions, assess and optimize all critical scaling groups to ensure reliable scale-out capacity when it matters most.