Can the automatic scaling provide the capability of vertical scaling?

Currently, the automatic scaling only provides the capability of horizontal scaling, namely, increase or decrease of number of instances, and cannot automatically adjust the instance configuration size. The user can automatically adjust the configuration size for the nodes in the scaling group.

What are the limits of quotas for the scaling group?

By default, one account can create a maximum of 10 scaling groups in the same region. Each scaling group can have a maximum of 20 nodes and 10 scaling rules. If you need a larger configuration, please submit a Ticket for application.

How long is the implementation period of the scaling rule according to plan?

After the scaling rule according to plan becomes valid, the scaling rule is generally implemented at the time point set by the rule. If the scaling rule fails to implement the scaling at the time point due to the implementation of other tasks, and other reasons, the rule keeps the "Trigger" status within 30 minutes. If the execution conditions can still not be reached after 30 minutes, the implementation of the rule fails.

How long is the time limit after the rule triggered by the alarm is triggered?

When the index of rule monitoring exceeds the threshold, the rule is always at the "Trigger" status. So the rule constantly tries to implement the scaling behavior until the monitoring data returns to normal or the data missing causes the rule abnormal.

What's the range of the acquisition time accuracy and delay of the monitoring data of the rule triggered by the alarm?

The accuracy of monitoring data is at the minute level, and the instance data are acquired once per minute. When the rule is just created, the automatic scaling registers the monitoring object in the cloud monitoring. So there may be a delay of 5-10 minutes. After a rule is created, the automatic scaling is subject to the data in Baidu AI Cloud monitoring BCM, and there may be a certain delay (about 3-6 minutes).

Does the number of nodes under the scaling group never exceed the maximum and minimum numbers of nodes?

The automatic scaling behaviors triggered by scaling rules never cause the actual number of nodes to exceed the boundary. However, the scaling behaviors manually initiated by the users may exceed maximum and minimum numbers of nodes. And the removal of abnormal nodes by the scaling group may also cause the actual number of nodes to be smaller than the minimum number of nodes. When finding the actual number of nodes exceeds the boundary, the scaling group initiates the behavior of scaling up to automatically adjust the number of nodes to the boundary.