View a markdown version of this page

DRHCSUS03-BP02 Monitor both workload resource utilization and Amazon EC2 instance consumption to maximize the use of AWS Outpost resources and improve sustainability - Data Residency and Hybrid Cloud Lens
Implementation guidance

DRHCSUS03-BP02 Monitor both workload resource utilization and Amazon EC2 instance consumption to maximize the use of AWS Outpost resources and improve sustainability

Portable workloads can be migrated from AWS Regions onto unused or underutilized compute resources on AWS Outposts to minimize energy consumption and maximize the use of fixed AWS Outposts resources.

Desired outcome: Outpost capacity is aligned to workload requirements and desired resiliency objectives. If, over time, an Outpost becomes underutilized, suitable workloads can be migrated from the Region to use the Outpost's capacity (while maintaining capacity for resiliency), reducing consumption in AWS Regions to support sustainability objectives.

Benefits of establishing this best practice: You can maximize utilization of your fixed, underutilized AWS Outposts compute resources by migrating portable workloads out of AWS Regions to minimize overall energy consumption.

Level of risk exposed if this best practice is not established: Medium

Implementation guidance

AWS Outposts provide a long-term, fixed, and finite pool of compute resources for addressing data residency and low latency use cases. Because AWS Outposts capacity is both fixed and purchased for terms spanning one to five years, it is critical that it be used to the fullest extent possible.

Use services such as Amazon CloudWatch metrics and AWS Compute Optimizer to monitor workload and instance metrics such as CPU utilization, memory utilization, network connections, and network throughput. Identify instances that have been over-provisioned and adjust as necessary to match instance type and size to the workload requirements. By matching instance type and size to workload requirements, you can free capacity on an Outpost to support workloads that would otherwise be deployed in the parent Region, which provides maximum value from the capacity already purchased on the Outpost.

When moving workloads to AWS Outposts, maintain spare capacity to absorb workloads if hardware in the Outpost fails. An Amazon EC2 server to spare ratio of eight to one is commonly used when planning for instance resiliency and potential hardware failures.