Steps involved in Autoscale configurations using metrics
Azure platform also gives you support to perform Autoscale configurations using metrics or CPU metrics. Taking this approach to Autoscale allows your website to scale regardless of any recurring or non-recurring schedules. To configure Autoscale based on the CPU metric you need to first set the Scale By Metric setting to CPU. This will activate the Instance Count and Target CPU controls in the portal that you can customize.The Instance Count and Target CPU settings are ranges of values, not a specific value. The way Azure interprets these settings is as follows:
- If the average CPU across all instances of your website are within the target CPU range, no action is taken.
- If the average CPU falls below the low-end of the target CPU range, Azure Autoscales your running instances down (once every two hours) until the number of running instances equals the low-end range defined in the Instance Count setting.
- If the average CPU exceeds the high-end of the target CPU range, Azure Autoscales your running instances up until the number of running instances equals the high-end range defined in the Instance Count setting.
New metrics are being added to the Autoscale feature that include Memory Percentage, Disk Queue Length, HTTP Queue Length, Data In, and Data Out. These metrics are only available in the new management portal at http://portal.azure.com. Figure 1 shows an example Autoscale configuration base on the CPU metric with an Instance Count between 3 and 8, and a Target CPU between 55 and 75 percent.
This article is a part of 70-533 Implementing Microsoft Azure Infrastructure Solutions Prep course.
More Articles included in this course are
Practice Tests in this course are