About Automatic Scaling

You can configure an Oracle Java Cloud Service instance to automatically scale a cluster in or out by defining an auto-scaling rule. These rules are based on CPU and/or memory utilization and determine when to add or remove nodes.

Topics:

How It Works

Learn to define a service and the conditions under which it should automatically add or delete a node.

You can define to a service the conditions under which it should automatically add or delete a node. These conditions are called rules (or a "policy") and are applicable to all the clusters in your service. The key component of the rules is the metric threshold that you set. You can choose for these metrics a percentage of CPU or memory utilization over a defined interval or the total gigabytes of memory consumed. Once the metric threshold is crossed, auto-scaling receives an alert and either adds or removes a single node, depending up the rule.

After that scaling operation succeeds, auto-scaling goes into a user-defined "cool down" period until the CPU utilization dips below the metric threshold or the cluster size reaches the user defined maximum/minimum cluster size. After cool down, if the alarm is still active (that is, if CPU utilization is still over—or below, depending on the scaling rules—the metric threshold), the service repeats the scaling operation until the CPU utilization dips below the metric threshold.

Prerequisites

Learn about the prerequisites to automatically scale a cluster in or out by defining an auto-scaling rule.

Before auto-scaling occurs, the system checks to ensure that the following prerequisites are met:

The rule is active.
The node is configured to handle the conditions of the rule.
For a scale out, the current cluster size must be smaller than the maximum cluster size defined in the rule
For a scale in, the current cluster size must be larger than the minimum cluster size defined in the rule.

What Are the Rule Components

Learn about the metric threshold, which consists of either the average, minimum, or maximum percentage of CPU usage,and others in the auto-scaling rule .

In addition to the metric threshold, which consists of either the average, minimum, or maximum percentage of CPU usage, the auto-scaling rule is composed of:

The scaling operation and, depending on that operation, maximum or minimum cluster size.
The number of consecutive times per a specific period the threshold must be crossed to trigger an alarm.
Whether the rule applies to any or all VM instances.
The duration of the cool-down period.