Cost optimization provides a recommended process for managing costs and describes strategies for major application types. The strategies vary depending on the application type and business architecture. This topic describes the cost optimization flow and typical scenarios.
Cost optimization flow
Based on FinOps principles, cost optimization in cloud-native scenarios involves three main steps:
You can set up cost centers and governance cycles based on your business needs.
Establish a cost baseline as needed.
Refer to the following ACK cost optimization overview flowchart to determine and execute an optimization plan.
Typical cost optimization scenarios
The following sections describe cost optimization scenarios for different stages of cloud-native adoption.
Infrastructure cloud-native transformation
This stage involves migrating a traditional IT architecture to the cloud or transforming it to be cloud-native. This process requires a major change to the infrastructure, making it a critical time to plan and manage costs. During this stage, focus on the following items:
Use the cost analysis feature of ACK to evaluate business system capacity. For more information, see Cost analysis feature.
Select instance types for the infrastructure. For more information, see Instance type selection guide.
Use savings plans to reduce your infrastructure costs. For more information, see What is a savings plan?.
Application cloud-native transformation
During the application cloud-native transformation, you can use Kubernetes features such as elastic scaling and mixed deployment to improve high availability. This allows containers, the standard deployment units, to automatically scale as service traffic fluctuates. This enables intelligent peak-load shifting, where resources are requested based on actual business demand.
Stable cloud-native services
After the cloud-native transformation is complete, you can create cost administration policies based on the dynamic changes of your services. Common scenarios include the following:
The business is cyclical.
Services show clear periodic fluctuations. For example, traffic might peak between 9 a.m. and 5 p.m. In this scenario, you can use the cost analysis feature to observe these patterns and then apply suitable cost optimization methods, such as elastic scaling. For more information, see Cost analysis feature.
Frequent launch and retirement of services
Frequent replacement of old services with new ones is common in many emerging business areas. In the emerging stage of a service, planning for application resource costs is a challenge. You can use the cost analysis feature to observe the resource usage of applications and the resource profiling feature to properly plan application configurations and specifications. For more information, see Cost analysis feature and Resource profiling.