# Optimizing your cost with rightsizing recommendations The rightsizing recommendations feature in Cost Explorer helps you identify cost-saving opportunities by downsizing or terminating instances in Amazon Elastic Compute Cloud (Amazon EC2). Rightsizing recommendations analyze your Amazon EC2 resources and usage to show opportunities for how you can lower your spending. You can see all of your underutilized Amazon EC2 instances across member accounts in a single view to immediately identify how much you can save. After you identify your recommendations, you can take action on the Amazon EC2 console. **Note** We recommend that you use Cost Optimization Hub to identify cost optimization opportunities. For full details, see [Cost Optimization Hub](https://docs.aws.amazon.com/cost-management/latest/userguide/cost-optimization-hub.html). Rightsizing recommendations doesn't support billing transfer. **Topics** + [ # Getting started with rightsizing recommendations ](rr-getting-started.md) + [ # Using your rightsizing recommendations ](rr-use.md) + [ # Sharing your rightsizing recommendations ](csv-details.md) + [ # Understanding rightsizing recommendations calculations ](understanding-rr-calc.md) + [ # Understanding reservations in Cost Explorer ](ce-ris.md) + [ # Accessing reservation recommendations ](ri-recommendations.md) # Getting started with rightsizing recommendations You can access your reservation recommendations and resource-based recommendations in the Billing and Cost Management console. After you enable the feature, it can take up to 24 hours to generate your recommendations. **To enable and access rightsizing recommendations** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. In the navigation pane, choose **Cost Management preferences**. 1. On the **Preferences** page, under **Rightsizing - *legacy*** in the **General** tab, select **Enable Rightsizing recommendations**. 1. Choose **Save preferences**. **Note** Only regular or a management account can enable rightsizing recommendations. After you enable the feature, both member and management account can access rightsizing recommendations unless the management account specifically prohibits member account access on the **settings** page. To improve the recommendation quality, AWS might use your published utilization metrics, such as disk or memory utilization, to improve our recommendation models and algorithms. All metrics are anonymized and aggregated before AWS uses them for model training. If you want to opt out of this experience and request that your metrics not be stored and used for model improvement, contact AWS Support. For more information, see [AWS Service Terms](https://aws.amazon.com/service-terms/). 1. To access rightsizing recommendations, in the navigation pane, under **Legacy pages**, choose **Rightsizing**. # Using your rightsizing recommendations You can see the following top-level key performance indicators (KPIs) in your rightsizing recommendations: + **Optimization opportunities** – The number of recommendations available based on your resources and usage + **Estimated monthly savings** – The sum of the projected monthly savings associated with each of the recommendations provided + **Estimated savings (%)** – The available savings relative to the direct instance costs (On-Demand) associated with the instances in the recommendation list **To filter your rightsizing recommendations** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. In the navigation pane, under **Legacy pages**, choose **Rightsizing**. 1. On the **Rightsizing recommendations** page, under **Recommendation parameters**, filter your recommendations by selecting any or all of the following check boxes: + Idle instances + Underutilized instances + Include Savings Plans and Reserved Instances 1. Under **Findings**, use the search bar to filter by the following parameters: + Account ID (option available from the management account) + Region + Cost allocation tag **To view your rightsizing recommendations details** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. In the navigation pane, under **Legacy pages**, choose **Rightsizing**. 1. On the **Rightsizing recommendations** page, under **Findings**, choose a recommendation to view the details. ## Enhancing your recommendations using CloudWatch metrics We can examine your memory utilization if you enable your Amazon CloudWatch agent. To enable memory utilization, see [Installing the CloudWatch Agent](https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/install-CloudWatch-Agent-on-EC2-Instance.html). **Important** When you create a CloudWatch configuration file, use the default namespace and default names for the collected metrics. For **InstanceID**, choose `append_Dimension`. Do not add additional dimensions for individual memory or disk metrics. Disk utilization is currently not examined. For Linux instances, choose `mem_used_percent` as your metric for your CloudWatch agent to collect. For Windows instances, choose `"% Committed Bytes In Use"`. For more information about the CloudWatch agent, see [Collecting Metrics and Logs from Amazon EC2 Instances and On-Premises Servers with the CloudWatch Agent](https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/Install-CloudWatch-Agent.html) in the *Amazon CloudWatch User Guide*. # Sharing your rightsizing recommendations You can download a rightsizing recommendations report in CSV format. **To download your recommendations** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. In the navigation pane, under **Legacy pages**, choose **Rightsizing**. 1. Under **Findings**, choose **Download CSV**. The following is a list of fields in the downloadable CSV file from the **Rightsizing recommendations** page. The fields are repeated if there are multiple rightsizing options available. The file also contains all of your relevant cost allocation tags. + **Account ID** – The AWS account ID that owns the instance that the recommendation is based off of. + **Account Name** – The name of the account that owns the instance that the recommendation is based off of. + **Instance ID** – The unique instance identifier. + **Instance Name** – The name you've given to the instance. + **Instance Type** – The instance family and size of the original instance. + **Instance Name** – The name you've given an instance. This field will show as blank if you haven't given the instance a name. + **OS** – The operating system or platform of the current instance. + **Region** – The AWS Region that the instance is running in. + **Running Hours** – The total number of running hours of the instance over the last 14 days. + **RI Hours** – The subset of the total running hours that are covered by an AWS reservation over the look-back period. + **OD Hours** – The subset of the total running hours that are On-Demand over the look-back period. + **SP Hours** – The subset of the total running hours that are covered by Savings Plans over the look-back period. + **CPU Utilization** – The maximum CPU utilization of the instance over the look-back period. + **Memory Utilization** – The maximum memory utilization of the instance over the look-back period (if available from the Amazon CloudWatch agent). + **Disk Utilization** – The maximum disk utilization of the instance over the look-back period (if available from the CloudWatch agent - currently not supported). + **Network Capacity** – The maximum network input/output operations per second capacity of the current instance. This isn't a measure of actual instance use or performance, only capacity. It's not considered in the recommendation. + **EBS Read Throughput** – The maximum number of read operations per second. + **EBS Write Throughput** – The maximum number of write operations per second. + **EBS Read Bandwidth** – The maximum volume of read KiB per second. + **EBS Write Bandwidth** – The maximum volume of write KiB per second. + **Recommended Action** – The recommended action, either modify or terminate the instance. + **Recommended Instance Type 1** – The instance family and size of the recommended instance type. For termination recommendations, this field is empty. + **Recommended Instance Type 1 Estimated Saving** – The projected savings based on the recommended action, instance type, associated rates, and your current Reserved Instance (RI) portfolio. + **Recommended Instance Type 1 Projected CPU** – The projected value of the CPU utilization based on utilization of current instance CPU and recommended instance specifications. + **Recommended Instance Type 1 Projected Memory** – The projected value of the memory utilization based on utilization of current instance memory and recommended instance specifications. + **Recommended Instance Type 1 Projected Disk** – The projected value of the disk utilization based on utilization of current instance disk and recommended instance specifications. + **Recommended Instance Type 1 Network Capacity** – The maximum network input/output operations per second capacity of the recommended instance. This isn't a measure of actual instance use or performance, only capacity. It's not considered in the recommendation. # Understanding rightsizing recommendations calculations This section provides an overview of the savings calculations that are used in your rightsizing recommendations algorithms. ## Consolidated billing family To identify all instances for all accounts in the consolidated billing family, rightsizing recommendations look at the usage for the last 14 days for each account. If the instance was stopped or terminated, we remove it from consideration. For all remaining instances, we call CloudWatch to get maximum CPU utilization data, memory utilization (if enabled), network in/out, local disk input/ output (I/O), and performance of attached EBS volumes for the last 14 days. This is to produce conservative recommendations, not to recommend instance modifications that could be detrimental to application performance or that could unexpectedly impact your performance. ## Determining if an instance is idle, underutilized, or neither We look at the maximum CPU utilization of the instance for the last 14 days to make one of the following assessments: + **Idle** – If the maximum CPU utilization is at or below 1%. A termination recommendation is generated, and savings are calculated. For more information, see [Savings calculation](#savings-calc). + **Underutilized** – If the maximum CPU utilization is above 1% and cost savings are available in modifying the instance type, a modification recommendation is generated. If the instance isn't idle or underutilized, we don't generate any recommendations. ## Generating modification recommendations Recommendations use a machine learning engine to identify the optimal Amazon EC2 instance types for a particular workload. Instance types include those that are a part of AWS Auto Scaling groups. The recommendations engine analyzes the configuration and resource usage of a workload to identify dozens of defining characteristics. For example, it can determine whether a workload is CPU-intensive or whether it exhibits a daily pattern. The recommendations engine analyzes these characteristics and identifies the hardware resources that the workload requires. Finally, it concludes how the workload would perform on various Amazon EC2 instances to make recommendations for the optimal AWS compute resources that the specific workload. ## Savings calculation We first examine the instance running in the last 14 days to identify whether it was partially or fully covered by an RI or Savings Plans, or running On-Demand. Another factor is whether the RI is size-flexible. The cost to run the instance is calculated based on the On-Demand hours and the rate of the instance type. For each recommendation, we calculate the cost to operate a new instance. We assume that a size-flexible RI covers the new instance in the same way as the previous instance if the new instance is within the same instance family. Estimated savings are calculated based on the number of On-Demand running hours and the difference in On-Demand rates. If the RI isn't size-flexible, or if the new instance is in a different instance family, the estimated savings calculation is based on whether the new instance had been running during the last 14 days as On-Demand. Cost Explorer only provides recommendations with an estimated savings greater than or equal to \$10. These recommendations are a subset of Compute Optimizer results. For more performance-based recommendations that might result in a cost increase, see [Compute Optimizer](https://aws.amazon.com/compute-optimizer/). You can choose to view saving with or without consideration for RI or Savings Plans discounts. Recommendations consider both discounts by default. Considering RI or Savings Plans discounts might result in some recommendations showing a savings value of \$10. To change this option, see [Using your rightsizing recommendations](rr-use.md). **Note** Rightsizing recommendations doesn't capture second-order effects of rightsizing, such as the resulting RI hour’s availability and how they will apply to other instances. Potential savings based on reallocation of the RI hours aren't included in the calculation. # Understanding reservations in Cost Explorer Balancing your reservation usage and your On-Demand instance or provisioned capacity usage can help you achieve better efficiency. To help, Cost Explorer provides tools that help you understand where your greatest reservation costs are and how you can potentially lower your costs. Cost Explorer provides you with an overview of your current reservations, shows your utilization and coverage, and calculates reservation recommendations that could save you money if you purchase them. ## Using your reservation reports You can use the **Reservations Overview** page in the Billing and Cost Management console to see how many reservations you have, how much your reservations are saving you compared to similar usage of On-Demand Instances, and how many of your reservations are expiring this month. Cost Explorer breaks down your reservations and savings by service and lists your potential savings; that is, the cost of On-Demand usage compared to what that usage could cost you with a reservation. To use your potential savings, see [Accessing reservation recommendations](ri-recommendations.md). ## Managing your reservation expiration alerts You can track your reservations and when those reservations expire in Cost Explorer. With reservation expiration alerts, you receive email alerts 7, 30, or 60 days in advance before your reservation expires. These alerts can be sent to up to 10 email recipients. You can also choose to be notified on the day that your reservation expires. Reservation expiration alerts are supported for Amazon EC2, Amazon RDS, Amazon Redshift, Amazon ElastiCache, and Amazon OpenSearch Service reservations. **To turn on reservation expiration alerts** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. Navigate to the **Overview** page under the **Reservations** section. 1. Choose **Manage alert subscriptions** in the upper right corner. 1. Select the check boxes for when you want to receive your alerts. 1. Enter email addresses for who you want to notify. You can have up to 10 email recipients. 1. Choose **Save**. AWS starts monitoring your reservation portfolio and sends alerts based on the preferences that you specify. # Accessing reservation recommendations If you enable Cost Explorer, you automatically get Amazon EC2, Amazon RDS, ElastiCache, OpenSearch Service, Amazon Redshift, Amazon MemoryDB, and Amazon DynamoDB purchase recommendations that could help reduce your costs. Reservations provide a discounted hourly rate (up to 75%) compared to On-Demand or provisioned capacity pricing. Cost Explorer generates your reservation recommendations using the following process: + Identifies your On-Demand instance or provisioned capacity usage for a service during a specific time period + Collects your usage into categories that are eligible for a reservation + Simulates every combination of reservation in each category of usage + Identifies the best number of each type of reservation to purchase to maximize your estimated savings For example, Cost Explorer automatically aggregates your Amazon EC2 Linux, shared tenancy, and c4 family usage in the US West (Oregon) Region and recommends that you buy size-flexible regional RIs to apply to the c4 family usage. Cost Explorer recommends the smallest size instance in an instance family. This makes it easier to purchase a size-flexible RI. Cost Explorer also shows the equal number of normalized units so that you can purchase any instance size that you want. For this example, your RI recommendation would be for `c4.large` because that is the smallest size instance in the c4 instance family. Cost Explorer recommendations are based on a single account or organization usage of the past seven, 30, or 60 days. Cost Explorer uses On-Demand instance usage during the selected look-back period to generate recommendations. All other usage in the look-back period that are covered by features such as RI, SPOT, and Savings Plans aren't included. Amazon EC2, ElastiCache, OpenSearch Service, Amazon Redshift, Amazon MemoryDB, and Amazon DynamoDB recommendations are for reservations scoped to Region, not Availability Zones, and your estimated savings reflects the application of those reservations to your usage. Amazon RDS recommendations are scoped to either Single-AZ or Multi-AZ RIs. Cost Explorer updates your recommendations at least once every 24 hours. **Note** Cost Explorer doesn't forecast your usage or take forecasts into account when recommending reservations. Instead, Cost Explorer assumes that your historical usage reflects your future usage when determining which reservations to recommend. Linked accounts can only see recommendations if they have the relevant permissions. Linked accounts need permissions to view Cost Explorer and permissions to view recommendations. For more information, see [Viewing reservation recommendations](#viewing-rex). **Topics** + [ ## RI recommendations for size-flexible RIs ](#rex-flex) + [ ## Viewing reservation recommendations ](#viewing-rex) + [ ## Understanding reservation recommendations ](#reading-rex) + [ ## Modifying reservation recommendations ](#modifying-rex) + [ ## Saving reservation recommendations ](#saving-rex) + [ ## Using reservation recommendations ](#using-rex) ## RI recommendations for size-flexible RIs Cost Explorer also considers the benefits of size-flexible regional RIs when generating your RI purchase recommendations. Size-flexible regional RIs help maximize your estimated savings across eligible instance families in your recommendations. AWS uses the concept of normalized units to compare the various sizes within an instance family. Cost Explorer uses the smallest normalization factor to represent the instance type that it recommends. For more information, see [Instance size flexibility](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/apply_ri.html#ri-instance-size-flexibility) in the *Amazon Elastic Compute Cloud User Guide*. For example, let’s say you own an EC2 RI for a `c4.8xlarge`. This RI applies to any usage of a `Linux/Unix c4` instance with shared tenancy in the same region as the RI, such as the following instances: + One `c4.8xlarge` instance + Two `c4.4xlarge` instances + Four `c4.2xlarge` instances + Sixteen `c4.large` instances It also includes combinations of EC2 usage, such as one `c4.4xlarge` and eight `c4.large` instances. If you own an RI that is smaller than the instance that you're running, you are charged the prorated, On-Demand price for the excess. This means that you could buy an RI for a `c4.4xlarge`, use a `c4.4xlarge` instance most of the time, but occasionally scale up to a `c4.8xlarge` instance. Some of your `c4.8xlarge` usage is covered by the purchased RI, and the rest is charged at On-Demand prices. For more information, see [How Reserved Instance discounts are applied](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/apply_ri.html) in the *Amazon Elastic Compute Cloud User Guide*. ## Viewing reservation recommendations Linked accounts need the following permissions to view recommendations: + `ViewBilling` + `ViewAccount` For more information, see [Using identity-based policies (IAM policies) for AWS Cost Management](billing-permissions-ref.md). **To view your reservation recommendations** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. In the navigation pane, under **Reservations**, choose **Recommendations**. 1. On the **Recommendations** page, under **Recommendation parameters**, choose the **Service** that you want recommendations for. ## Understanding reservation recommendations The **Reservations Recommendations** page shows you your estimated potential savings, your reservation purchase recommendations, and the parameters that Cost Explorer used to create your recommendations. You can change the parameters to get recommendations that might match your use case more closely. The **Recommendations** page shows you the following three numbers: + **Total purchase recommendations** – The number of different reservation purchase options Cost Explorer has found for you. + **Estimated monthly savings** – How much Cost Explorer calculates you could save by purchasing the recommended reservations. + **Estimated savings vs. On-Demand rates** – Your estimated savings as a percentage of your current costs. These numbers provide you with a rough estimate of how much you could potentially save by buying more reservations. You can recalculate these numbers for a different use case using the following **Recommendation parameters**: + **Term** – The duration for which you want recommendations. + **Offering class** – Whether you want recommendations for a standard or convertible reservation. + **Payment option** – Whether you want to pay for recommendations upfront. + **Based on the past** – The number of days of previous usage that you want your recommendations to take into account. At the bottom of the page are tabs with some of your savings estimates. The **All accounts** tab enables you to see the recommendations based on the combined usage across your entire organization, and the **Individual accounts** tab enables you to see recommendations that Cost Explorer generated on a per-linked-account basis. The table on each tab shows the different purchase recommendations and details about the recommendations. If you want to see the usage that Cost Explorer based a recommendation on, choose the **View associated usage** link in the recommendation details. This takes you to a report that shows the exact parameters that Cost Explorer used to generate your recommendation. The report also shows your costs and associated usage grouped by **Purchase option**, so that you can view the On-Demand Instance usage that your recommendation is based on. **Note** Recommendations that Cost Explorer bases on an individual linked account consider all usage by that linked account, including any RIs used by that linked account. This includes RIs shared by another linked account. The recommendations don't assume that an RI will be shared with the linked account in the future. When you use billing transfer as a bill source account, your estimated savings in reservation recommendations might not match the effective savings data in your AWS Cost and Usage Report, Cost Explorer, and Bills page. This is because reservation recommendations don't use custom data generated by Billing Conductor. You can sort your recommendations by **Monthly estimated savings**, **Upfront RI cost**, **Purchase recommendation**, or **Instance type**. ## Modifying reservation recommendations You can change the information that Cost Explorer uses when it creates your recommendations, and you can also change the types of recommendations that you want. This allows you to see recommendations for the reservations that work best for you, such as All upfront reservations with a one-year term, based on your last 30 days of usage. **Note** Instead of forecasting your future usage, Cost Explorer assumes that your future usage is the same as your previous usage. Cost Explorer also assumes that you are renewing any expiring reservations. **To modify your reservation recommendations** 1. Open the Billing and Cost Management console at [https://console.aws.amazon.com/costmanagement/](https://console.aws.amazon.com/costmanagement/). 1. In the navigation pane, under **Reservations**, choose **Recommendations**. 1. On the **Recommendations** page, under **Recommendation parameters**, choose the **Service** that you want recommendations for. 1. Choose the relevant **Term**. 1. Choose the relevant **Offering class**. 1. Choose the relevant **Payment option**. 1. For **Based on the past**, select how many days of usage that you want your reservation recommendations to be based on. 1. Choose either **All accounts** or **Individual accounts** to see recommendations based either on your organization-wide usage or on all of your linked accounts based on their individual account usage. ## Saving reservation recommendations You can save reservation recommendations as a CSV file. **To save your reservation recommendations** 1. On the **Reservations Recommendations** page, under **Recommendation parameters**, choose the **Service** that you want recommendations for and update any parameters you want to change. 1. Under **Recommended actions**, choose **Download CSV**. The CSV file contains the following columns. **Reservation recommendations CSV columns** | Column name | Service | Column explanation | | --- | --- | --- | | Account ID | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The account associated with your recommendation. | | Availability zone | Amazon RDS | The availability zone of the instances used to generate a recommendation. | | Average hourly normalized unit usage in historical period | Amazon EC2, RDS, MemoryDB | The average number of normalized units used per hour over the period chosen for generating recommendations. | | Average hourly usage in historical period | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB | The average number of instance hours used per hour over the period chosen for generating recommendations. | | Average number of capacity units used per hour in the selected historical period | Amazon DynamoDB | The average number of provisioned capacity units used per hour over the period chosen for generating recommendations. | | Break even months | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The estimated length of time before you recoup your upfront costs for this set of recommended reservations. | | Cache engine | Amazon ElastiCache | The kind of engine that the recommended ElastiCache reserved node runs, such as Redis or Memcheched. | | Capacity unit type | Amazon DynamoDB | The type of capacity unit for the recommendation. Read capacity units are used for operations that retrieve data from a table. Write capacity units are used for operations that insert, update, or delete data in a table. | | Database edition | Amazon RDS | The edition of the database engine that the recommended RDS reserved instance runs. | | Database engine | Amazon RDS | The kind of engine that the recommended RDS reserved instance runs, such as Aurora MySQL or MariaDB. | | Deployment model | Amazon RDS | Whether your recommended reserved instance is for RDS Custom. If the value is "Custom", you must to purchase the recommended reserved instances for RDS Custom to see potential savings. | | Deployment option | Amazon RDS | Whether your reserved instance is for an RDS instance in a single Availability Zone or an RDS instance with a backup in another Availability Zone. | | Estimated savings | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The estimated savings of the recommended reservations. | | Expected utilization | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | How much of the recommended reservations Cost Explorer estimates you will use. | | Instance type | Amazon EC2, RDS, OpenSearch Service | The type of instance that the recommendation is generated for (for example, m4.large or t2.nano). For size-flexible recommendations, Cost Explorer aggregates all usage in a organization (for example, the m4 family) and shows a recommendation for the smallest reserved instance type that is available for purchase (for example, m4.large). | | Max hourly normalized unit usage in historical period | Amazon EC2, RDS, MemoryDB | The maximum number of normalized units used in an hour over the period chosen for generating recommendations. | | Max hourly usage in historical period | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB | The maximum number of instance hours used in an hour over the period chosen for generating recommendations. | | Maximum number of capacity units used per hour in the selected historical period | Amazon DynamoDB | The maximum number of provisioned capacity units used in an hour over the period chosen for generating recommendations. | | Min hourly normalized unit usage in historical period | Amazon EC2, RDS, MemoryDB | The minimum number of normalized units used in an hour over the period chosen for generating recommendations. | | Min hourly usage in historical period | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB | The minimum number of instance hours used in an hour over the period chosen for generating recommendations. | | Minimum number of capacity units used per hour in the selected historical period | Amazon DynamoDB | The minimum number of provisioned capacity units used in an hour over the period chosen for generating recommendations. | | Node type | Amazon ElastiCache, Redshift, MemoryDB | The type of node that the recommendation is generated for, such as ds2.xlarge. | | Normalized hours to purchase | Amazon EC2, RDS, MemoryDB | How many normalized units that Cost Explorer recommends that you buy. | | Number of instances to purchase | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB | How many reservations Cost Explorer recommends that you buy. | | Offering class | Amazon EC2 | The offering class associated with the recommendation. | | Payment option | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The recommended payment option for the recommendation. | | Platform | Amazon EC2 | The operating system and license model for the recommended reserved instance type. | | Recommended number of capacity units to purchase | Amazon DynamoDB | How many reserved capacity units Cost Explorer recommends that you buy. | | Recommendation date | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The date that Cost Explorer generated your recommendation. | | Recurring monthly cost | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The recurring monthly cost of the recommended reservations. | | Region | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The Region used to generate a recommendation. You must purchase the recommended reservations in the recommended Region to see potential savings. | | Size flexible | Amazon EC2, RDS, MemoryDB | Whether a recommended reservation is size-flexible. | | Tenancy | Amazon EC2 | The tenancy for the recommendation. Valid values are shared or dedicated. | | Term | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The recommended term length for the recommendation. | | Upfront cost | Amazon EC2, RDS, Redshift, ElastiCache, OpenSearch Service, MemoryDB, DynamoDB | The upfront cost associated with the recommendation. | ## Using reservation recommendations To purchase the recommended reservations, go to the purchase page in a service console. You can also save a CSV file of your recommendations and purchase the reservations at a later date. **To use Amazon Elastic Compute Cloud recommendations** 1. On the **Reserved Instance Recommendations** page, choose [Amazon EC2 RI Purchase Console](https://console.aws.amazon.com/ec2/v2/home?region=us-east-1#ReservedInstances:sort=reservedInstancesId). 1. Purchase your RIs by following the instructions at [Buy Reserved Instances for Amazon EC2](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ri-market-concepts-buying.html) in the *Amazon Elastic Compute Cloud User Guide*. **To use Amazon Relational Database Service recommendations** 1. On the **Reserved instances** page in the Amazon RDS console, choose **Purchase Reserved DB Instance**. 1. Purchase your reservations by following the instructions at [Purchasing reserved DB instances for Amazon RDS](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_WorkingWithReservedDBInstances.WorkingWith.html) in the *Amazon RDS User Guide*. **To use Amazon Redshift recommendations** 1. On the **Reserved nodes** page in the Amazon Redshift console, choose **Purchase reserved nodes**. 1. Purchase your reservations by following the instructions at [Purchasing a reserved node](https://docs.aws.amazon.com/redshift/latest/mgmt/purchase-reserved-node-offering-console.html) in the *Amazon Redshift Management Guide*. **To use Amazon OpenSearch Service recommendations** 1. On the **Reserved Instance Leases** page in the OpenSearch Service console, choose **Order Reserved Instance**. 1. Purchase your reservations by following the instructions at [ Reserved Instances in Amazon OpenSearch Service](https://docs.aws.amazon.com/opensearch-service/latest/developerguide/ri.html) in the *Amazon OpenSearch Service Developer Guide*. **To use Amazon ElastiCache recommendations** 1. On the **Reserved Nodes** page in the ElastiCache console, choose **Purchase reserved nodes**. 1. Purchase your reservations by following the instructions at [Purchasing a reserved node](https://docs.aws.amazon.com/AmazonElastiCache/latest/UserGuide/CacheNodes.Reserved.html#reserved-nodes-purchasing) in the *Amazon ElastiCache User Guide*. **To use Amazon MemoryDB recommendations** 1. On the **Reserved nodes** page in the MemoryDB console, choose **Purchase reserved nodes**. 1. Purchase your reservations by following the instructions at [Working with reserved nodes](https://docs.aws.amazon.com/memorydb/latest/devguide/nodes.reservednodes.html) in the *Amazon MemoryDB Developer Guide*. **To use Amazon DynamoDB recommendations** 1. On the **Reserved capacity** page in the DynamoDB console, choose **Purchase reserved capacity**. 1. Purchase your reserved capacity by following the instructions at [Reserved capacity](https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/reserved-capacity.html) in the *Amazon DynamoDB Developer Guide*.