Automatic Vertical Scaling

Automatic Vertical Scaling

Automatic vertical scaling dynamically adjusts the amount of CPU and RAM available to a container according to its current workload. This helps applications receive additional resources during traffic peaks while avoiding unnecessary resource allocation during periods of low activity.

In this article

Overview

Vertical scaling changes the resource capacity of an existing container without adding or removing nodes. The platform can automatically increase or decrease the available resources within the limits configured for that container.

Resources are allocated in cloudlets. Each cloudlet represents a predefined amount of CPU and RAM. The exact cloudlet size may depend on the hosting provider.

✓

Automatic resource allocation

The container receives additional resources only when they are required and releases unused capacity when the workload decreases.

How Automatic Vertical Scaling Works

The platform continuously monitors CPU and RAM consumption for each container. Resource allocation is adjusted automatically between the configured minimum and maximum limits.

Low workload The container operates using the reserved resources and only the dynamic resources that are currently required.
Increased workload Additional dynamic cloudlets are allocated automatically, up to the configured maximum limit.
Reduced workload Unused dynamic resources are released automatically after demand decreases.
Maximum limit reached The container cannot receive resources beyond the configured scaling limit unless the topology settings are changed.
!

Maximum resources are not reserved permanently

The upper scaling limit defines the maximum capacity available to the container. Dynamic resources within this limit are allocated only when the application actually needs them.

Reserved and Dynamic Cloudlets

Reserved Cloudlets Resources that are permanently assigned to the container. They remain available regardless of the current workload and are normally charged at a lower fixed rate.
Dynamic Cloudlets Additional resources that are allocated automatically when required. Billing normally applies only to the dynamic resources actually consumed.
Scaling Limit The total maximum number of reserved and dynamic cloudlets that the container can use.

A balanced configuration normally includes enough reserved cloudlets to support the application’s regular workload and sufficient dynamic cloudlets to absorb temporary usage peaks.

Configure Vertical Scaling

1

Open the environment topology wizard by clicking Change Environment Topology for the required environment.

2

Select the required node layer, such as an application server, database, load balancer, cache, or another container type.

3

Adjust the vertical scaling sliders for the selected node:

  • Set the number of Reserved Cloudlets.
  • Set the total Scaling Limit.
  • Review the corresponding CPU and RAM capacity.
4

Repeat the configuration for other node layers when required and click Apply to save the topology changes.

✓

No manual restart is normally required

Vertical scaling is designed to adjust available resources without replacing the container. The exact behavior may vary depending on the node type and hosting-provider configuration.

Configure Load Triggers

Load alerts can be configured to notify users or execute actions when resource consumption crosses a specified threshold. Open the environment settings and select Monitoring > Load Alerts.

Create a new trigger and configure the required parameters:

Node Layer Select the environment layer that should be monitored.
Resource Type Select CPU, memory, network, disk, or another available monitoring metric.
Condition Define whether the trigger activates when usage rises above or falls below the threshold.
Threshold Specify the resource-usage percentage or measurement that activates the trigger.
Duration Set how long the condition must remain true before the trigger is activated.
Action Select the notification or platform action to perform after the condition is met.
!

Load alerts do not replace the scaling limit

Automatic vertical scaling operates within the resource limits configured in the topology wizard. Alerts help monitor usage but cannot allocate resources above the defined maximum.

Configure Notifications

Notifications can help identify containers that are frequently approaching their maximum resource limit. Depending on the platform configuration, alerts may be delivered through email, dashboard messages, or supported integration channels.

High-resource warning Notify administrators when CPU or memory consumption remains above a defined level.
Maximum-limit warning Notify administrators when a container repeatedly reaches or remains close to its scaling limit.
Recovery notification Notify administrators when consumption returns below the defined threshold.

Monitor Resource Usage

Use the environment statistics and monitoring charts to review resource consumption over time. Monitoring helps determine whether the current reserved resources and scaling limit are appropriate.

CPU Usage Review processor consumption and identify sustained high-load periods.
RAM Usage Check whether memory consumption regularly approaches the configured maximum.
Cloudlet Usage Compare the number of consumed cloudlets with the reserved and maximum limits.
Historical Trends Use historical data to distinguish temporary spikes from permanent growth in application demand.

If the container frequently reaches its upper limit, increase the scaling limit or optimize the application. If the baseline usage is consistently high, increasing the reserved cloudlets may provide a more suitable and predictable configuration.

Recommendations and Limitations

Set a realistic baseline Reserve enough resources for the application’s normal workload instead of relying entirely on dynamic allocation.
Allow capacity for peaks Set the maximum limit high enough to handle expected traffic or processing spikes.
Monitor regularly Review usage charts and load alerts to identify whether limits should be adjusted.
Check application limits Application-level restrictions, such as JVM heap settings, PHP memory limits, or database configuration, may prevent software from using all available container memory.
Use horizontal scaling when needed Vertical scaling increases resources within one container. Use horizontal scaling when the application requires additional node instances, high availability, or distributed load handling.
Observe provider limits The maximum cloudlets, available resources, and billing model depend on the hosting provider and account type.
✓

Expected result

The selected container automatically receives additional CPU and RAM when workload increases and releases unused dynamic capacity when demand decreases, while remaining within the configured resource limit.

Related Documentation