Reshaping

Reshaping analyzes your instance level usage & utilization data, scanning thousands of potential AWS machine architectures to find hidden efficiencies you can gain back by reducing your near-idle machine resources.

Intelligently integrated into your FinOps spending data (including effective savings rates, and RI footprint). Customizable via your account preferences on provisioning & headroom.

Understanding Home Dashboard

Reshaping will show one service at a time, to toggle between services (such as Ec2, EBS, etc) use the toggle on the top right.

When you change what service is showing the suggestion box, at the bottom will change. You can click each line item for additional details, suggestion particulars & usage data.

Changing Your Reshaping Preferences

You can adjust the account preferences for how the machine learning model analyzes your machines and recommends improvements. To do this, click the gear icon on the reshaping homepage. Please note that changes will take 24 hours to appear in the output data as the analysis re-runs in the background. It is recommended to change preferences infrequently, ideally once a month or quarter, to avoid confusion.

Lookback Period is the period of historical data Reshaping will look at, before suggesting the new architecture will fit.

CPU Threshold will change what percentage of CPU usage spokes that the model will ignore, when it comes to suggesting new architecture. The lower the threshold, the more conservative the suggestions will be.

CPU Headroom will dictate how much CPU headroom you'd like the new architecture suggestions to keep. The more headroom, the less aggresive the savings suggestion.

Memory Headroom will dictate how much Memory headroom you'd like the new architecture suggestions to keep. The more headroom, the less aggresive the savings suggestion.

Cohorts, sharing & collaboration

Reshaping suggestions will show if any machines live within accounts that are mapped to a cohort. You can filter by cohort using filters.

You can select one or many reshaping suggestions, and push the suggestion to create a new Jira ticket, slack message or invite. The more collaboration the better.