Get recommendations for buying Reserved Instances and Savings Plans
The Reserved Instance Planner is one of our most popular features. It helps you to maximize your AWS investment by analyzing your usage to identify opportunities for savings by using Reserved Instances.
In this article, we'll cover:
- Reserved Instance fundamentals
- Generating recommendations
- Purchase recommendations
- Modification recommendations
- Underutilized Reserved Instances
- Missing Reservations
We just launched a new version of the RI Planner. This document will be updated to reflect the new look and feel.
Reserved Instance fundamentals
RI Planner does not support views that contain account group mappings. In AWS, Azure, and any other provider, tag mappings and potentially business mappings contain rules that restrict a user to only a portion of an account. To protect the privacy of customers, we only allow users with appropriate permissions to view the Reservation page, as Reservations deal with entire accounts. This is a limitation with data and how it interacts with view filtering.
Before we dive into the Reserved Instance Planner, take a moment to watch this video. It provides a quick refresher on RI fundamentals.
Video: https://cloudability.zendesk.com?wvideo=ul5uhscalr
Generating recommendations
The first step of using the Reserved Instance Planner is to select the account for which you want recommendations. While you can choose a specific linked account, we usually recommend using the master payer account. Doing this provides the opportunity of having the reservations applied across all the linked accounts grouped underneath it and maximizing your savings.
When you select a specific linked account, the recommendation will ignore any discounts purchased in the management (payer) account or other accounts, and will provide a recommendation based on the uncovered usage in the specified account.
Next, select the time frame you want to run your recommendations against.
The RI Planner defaults to a recent window of time based on the current date, but you should select the time period that is most representative of the current state of your infrastructure. For example, if today's date was August 18th, you could choose a 2-week period in the previous month.
It's important to note that the RI Planner will exclude All Upfront types by default in favor of Partial Upfront and No Upfront types.
On the surface, it may seem like All Upfront RI types will offer the most value. However, from our experience, there is more value in selecting a Partial Upfront type when making your RI purchasing decisions. The planner will then generate a list of RI recommendations based on buy or modify actions for each class of instance type, OS, and Availability Zone. Prioritizing making the RI modifications to get some quick cost savings wins before proceeding with your RI buys.

Buy recommendations
The first section of the Planner's recommendation list is the Buy
section. For each class of instance type, OS, and Availability Zone, Cloudability will tell you how many RIs of the selected type you
currently have, how many you need to maximize your coverage, the upfront cost, and the associated
savings. As a best practice, rather than purchasing all the recommended RIs, start with a small and
uncontroversial RI purchase using the AWS Partial Upfront option and plan
to buy iteratively on a schedule that you hold sacred .
To get more information about how the recommendation is calculated, as well as financial
breakeven points, mouse over the > arrow and click the
Details button that appears.
This will bring up a second screen with more details on that specific recommendation.
The graph will show you how many of this type of instance you're running each hour of the
reporting period. Below, you’ll see the number of months it will take to break even on the
reservation, as well as a breakdown of the financials.
Modification recommendations
Below the Buy section of the Planner's recommendations, you'll find
the Modify section. This will propose modifications to reservations which
you currently own but are not being fully utilized based on your current usage but could be if
reallocated.
On the left, you'll see a list of currently owned reservations that should be modified. On the right, you'll see a list of new reservations that the Planner recommends you modify from your current ones.
The Planner may recommend that you modify the size or Availability Zone of your reservations depending on where your infrastructure has the most need. You may notice the Units column. These units are calculated according to the normalization factor of the reservations listed.
You can modify any number of reservations into any number of different reservations, so long as the number of total units stays constant.
Underutilized Reserved Instances
At the bottom of its recommendations, the Reserved Instance Planner features a section of any underutilized RIs you currently own. The reservations listed here either can’t be modified or wouldn’t benefit from it. However, having this information provides an opportunity to take advantage of that investment.
Underutilized RIs can be put on a list and socialized within your organization (email, shared doc, etc.). So, the next time someone needs to spin up an AWS resource, encourage them to reference your list of unused RIs and use an instance of a matching type.
Missing Reservations
The RI tool can only see RIs that the user's account can see. An Amazon consolidated billing parent account cannot see the activity of the child account that occurred before it entered into the consolidated billing relationship. This means that the parent account cannot see reserved instances belonging to the child that were purchased before the consolidated billing relationship began.
The suggested workaround is to add an IAM credential for the child account to Cloudability , in addition to the consolidated billing parent account. The RI Tool will then be able to see the reservations that were purchased by the child account. Cloudability will automatically de-duplicate any overlapping data.