All Documents
Current Document

Content is empty

If you don't find the content you expect, please try another search term

Documentation

Monitor metrics

Last updated:2021-05-21 17:42:57

Monitor metrics

Metrics are time series data about the performance of your system. Auto Scaling uses the cloud platform monitoring capability to retrieve metric data. You can determine whether your system runs as expected based on these metrics. If the metrics exceed thresholds, Auto Scaling executes scale-out or scale-in activities.

Auto Scaling supports the following metrics:

  • CPU usage

  • Memory usage

Each metric supports the following dimensions:

  • Maximum value

  • Minimum value

  • Average value

Aggregate metric data

Auto Scaling monitors KEC instances in a scaling group, and collects metric data from these KEC instances in each cycle. Auto Scaling aggregates the metric data and triggers scaling activities based on your policies.

Basically, Auto Scaling collects the value of the specified metric from each KEC instance every 1 minute in each cycle. If the obtained values meet the specified condition in multiple consecutive cycles, Auto Scaling triggers a scaling activity. The number of consecutive cycles can be customized.

Example

A scaling group contains three KEC instances. An alarm-triggered policy is defined as follows: A scaling activity is triggered if the maximum, minimum, or average value of CPU usage exceeds 50% in a cycle, which is set to 5 minutes, and this occurs in three consecutive cycles.

Auto Scaling collects metric data, checks whether the data meets the condition specified in the alarm-triggered policy, and takes actions accordingly. The procedure is as follows:

  1. Auto Scaling collects the CPU usage of each KEC instance every 1 minute. In a cycle, which is set to 5 minutes, 15 CPU usage values are collected.

  2. Auto Scaling checks whether the maximum, minimum, or average value among the 15 values meets the specified condition.

    • Maximum value: If the maximum value among the 15 values exceeds 50%, the condition is met in this cycle.

    • Minimum value: If the minimum value among the 15 values exceeds 50%, the condition is met in this cycle.

    • Average value: If the average value among the 15 values exceeds 50%, the condition is met in this cycle.
  3. If the condition is met in three consecutive cycles, which are 15 minutes, Auto Scaling triggers a scaling activity.
On this page
Pure ModeNormal Mode

Pure Mode

Click to preview the document content in full screen
Feedback