You can use Instance Template to create a scaling group. The instance template pre-defines the component information of individual instances comprising the scaling group. See Instance Template Console Guide for details.
Shows currently-active scaling groups. On the View List screen, status of each scaling group can be found.
|CREATE||Create a scaling group|
|UPDATE||Change a scaling group
Changes to resources owned by a scaling group
|HANDOVER||Change a scaling group owner|
|SUSPEND||Stop a scaling group|
|RESUME||Start a scaling group|
|DELETE||Delete a scaling group|
|IN_PROGRESS||Action in progress|
[Note] A scaling group in a failed action may not operate normally. Please contact the Customer Center.
Following items can be defined in a scaling group.
|Setting||Name||The name of the scaling group, alphabet characters, '-', '.' and numbers, within 20 characters|
|Instance Template||The instance template to be used by a scaling group|
|Minimum Instance||The number of instances to be maintained at the minimum while a scaling group is activated|
|Maximum Instance||The number of maximum instances allowed while a scaling group is activated|
|Running Instance||The number of instances created when a scaling group is activated for the first time|
|Scale-out/in policy||Condition||Initiating conditions for scale-out/in policy |
Specify performance indicators, reference values, and continued time
|Conditional Operator||Operators to be applied between initiating conditions
With and, policy is initiated when all conditions are satisfied
With or, policy is initiated when only one of the conditions is met
|Adjust instance||Number of instances to be created or deleted when a policy is initiated.|
|Cooldown Period||Time to wait until a policy is initiated again after previous initiation
If cooldown period has not passed, policy cannot be initiated even if conditions are met.
|Auto healing policy||Auto healing||Whether to use auto healing policy|
|Load Balancer||Selected Load Balancer||The load balancer that a created instance is to be connected with.|
|Additional policy||Deploy linkage||Whether to use the auto distribution feature using deploy service when scale-out|
[Note] With auto healing policies, the failure of each individual instance is handled by deleting the instance and replacing it with a new one. If an instance's performance metrics are not collected during a continuous 3-minute period, it is determined as an error and auto healing will proceed. Auto healing occurs regardless of the cooldown period.
[Note] Enabling the Deploy linkage option when creating scaling groups allows users to use the Deploy service to automatically deploy their applications as new instances are created. For more information, see Deploy Guide. Deploy linkage feature is currently provided only in Korea (Pangyo, Pyeongchon) and Japan(Tokyo) regions as of July, 2021. When linking with Deploy, user scripts with unicode characters do not work.
You can attach a load balancer to a scaling group, remove or replace the already attached load balancer. You must create the load balancer to connect in advance.
[Note] Even if you add a listener to an already attached load balancer, instances in the current scaling group are not automatically connected to a new listner. If you need to connect to the new listener, you must disconnect and reconnect to the load balancer.
[Caution] To attach a load balancer to a scaling group, Infrastructure Load Balancer ADMIN or Infrastructure ADMIN permissions are required. When you change the load balancer, the existing instance is deleted and a new instance is created.
Selects a desired scaling group from the scaling group list and pauses it. The paused scaling group can be restarted using the 'Start Scaling Group' button.
[Note] The status of a paused scaling group is displayed in yellow. All control features including controlling the number of instances, changing scaling group policy, changing load balancer, creating scheduled tasks are limited for a paused scaling group. As the affiliated instances stop, statistics graphs are unavailable.
If you select the owner you want to change, the scaling groups owned by the selected owner will be displayed. Select the scaling group whose owner you want to change to yourself. After the change, you can manage the scaling group with the key pair selected when changing the owner.
[Note] When creating a scaling group, the owner is set to the user who requested creation.
[Caution] Instances created after the owner change can be accessed with the key pair selected when the owner is changed, but existing instances created before the change must still be accessed with the key pair before the change. Therefore, key pair files should be well managed at the user level.
Select a scaling group from the list of scaling groups and check its details.
Edit on details screen, to modify attributes of the scaling group. By modifying the scaling group, instance templates in use or minimum/maximum/running instances can be changed.
Select a scaling group from the list of scaling groups and check its scaling policy.
Edit on policy details, to modify scaling policy. Or, click
Execute in scale up/down policy to initiate the policy by force.
Select a scaling group from the list of scaling groups, and check scheduled tasks.
With task scheduling, the number of minimum/maximum/running instances of a scaling group at a specific time can be adjusted.
Scheduled tasks can be set for one-time or periodic execution.
Items as follows are required to create a scheduled task:
|Name||Name of a scheduled task|
|Change Items||Attributes of a scaling group to be changed by a scheduled task
Select one out of the minimum/maximum/running instances
|Value||New value of an attribute specified in change items
Modify the attribute selected from
|Repeat||Whether to repeat a scheduled task
Select either once or Cron expression
|Cron Expression||Activated when Cron expression is selected for
|Start Time||Activation time for a scheduled task
When Cron expression is selected for
|End Time||Closing time for a scheduled task
Activated when Cron expression is selected for
[Note] The Cron Expression is applied to show execution time/cycle of a scheduled task.
The Cron expression is comprised of five items, each of which is divided by space characters and it means as follows:
Item Range Allowed Available Special Characters Minute 0-59
For each item, use numbers or special characters to specify execution time. For correct grammar, refer to the following:
Special Characters Meaning * All Hours - Range , Specific Time / Increase Volume L Last Time W Closest Weekday: available only on Day items # The Nth Day: available only on Day items
The Cron expression is used like in the following examples.
0 10 * * *: Executes at every 10:00
0/20 15 * * *: Executes in every 20 minutes from 15:00, like 15:00, 15:20, and 15:40
0 12-15 * * *: Executes at every 12:00, 13:00, 14:00, and 15:00
0 0 15 6,7,8 *: Executes at 00:00 on the 15th of June, July, and August
0 0 L * *: Executes at 00:00 on the last day of every month
0 9 25W * *: Executes at 09:00 on a closest weekday from the 25th day of every month
0 9 * * 3#2: Executes at 09:00 on the second Thursday of every month
[Caution] Start time of a scheduled task can be specified only after three minutes from the current time. If a scaling group is now changing, scheduled task may be delayed.
Select a scaling group from the list and check the list of created instances.
[Caution] Instances that a scaling group created are also exposed on the list of instance products. However, user cannot control them.
Select a scaling group from the list and check its statistical graph.
[Note] The statistical graph shows the average usage of instances that belong to a scaling group during the last 1 day. You can also identify the point of time and cause of increase or decrease.
Statistical graphs provide statistics on system resources, as follows:
|System Resources||Provided Statistical Data|
|CPU Usage||Average CPU usage of all instances that belong to a scaling group|
|Memory Usage||Average memory usage of all instances that belong to a scaling group|
|Disk Transfer Rate||Average data volume of read/write disk data per minute for all instances that belong to a scaling group|
|Network Transfer Rate||Average data volume of send/receive network per minute for all instances that belong to a scaling group|
Each graph can be extended up to 5-minute intervals, providing 1-minute statistical data at the minimum.
[Note] Statistics require data of every instance of a scaling group, but it is impossible to collect all instance data at once. And that may cause some missing data during the first few minutes. Statistics are calculated based only on those instances which include collected data; hence, with further data added at later stage, values may change. Now, after all data is collected for every instance, you can find the same values at all times.