Savings Plan Compute Overview

The Workspace Savings Plan Compute – Overview aims to structure a technical and financial analysis to:

• Size the ideal hourly commitment for a Compute Savings Plan

• Reduce the risk of underutilization (undercommitment and overcommitment)

• Evaluate the historical stability of eligible consumption

• Understand the organizational concentration of spending

• Project financial impact based on real data

The focus is not just on percentage savings, but on financial predictability and commitment governance.

Data Source

• Data Connector: AWS Billing (Cost and Usage Report – CUR)

• Dataset: Dedicated workspace dataset

• Base Granularity: Billing line (line item)

• Currency: According to Context configuration

---

View - Daily Cost

Overview

This analysis aims to evaluate the volatility, trend, and daily consumption patterns of resources eligible for the Compute Savings Plan. It allows for the identification of the operational baseline stability, assisting in defining purchase commitments that are resilient to day-to-day variations.

What the Analysis Considers

To reflect the genuine operational behavior of the environment, the analysis is based on:

• Qualified Consumption: Only recurring consumption effectively eligible for the plan.

• Data Purity: Exclusion of non-recurring financial effects (credits, refunds, or adjustments).

• Real Load: Focus on active computational load effectively executed during the selected period.

• Temporal Consistency: Analysis based on the full period defined in the search filters.

---

Data Structure and Grouping

Dimensions (Column)

• Usage Start Date (lineitem/usagestartdate): Represents the beginning of the consumption period recorded in the CUR.

• Granularity: Configured with daily grouping, allowing for precise temporal analysis of stability or seasonality of eligible costs.

Metrics (Data)

• Unblended Cost (lineitem/unblendedcost): Indicates the gross cost of consumption before any application of contractual discounts or reservation amortizations.

• Aggregation (SUM): Consolidated sum of eligible costs per day, with the total enabled for checking global financial volume.

---

Selection Filters and Eligibility

The precision of this analysis depends on rigorous filters that isolate "committable" consumption:

• Line Item Type: lineitem/lineitemtype IS Usage

  • Restricts the analysis to actual consumption. Excludes credits, support fees, and Savings Plans/RI lines already applied.

• Compute Services: product/productname IS Amazon Elastic Compute Cloud, AWS Lambda, Amazon ECS, and Amazon EKS.

• Usage Type Refinement (usagetype):

  • Contains: Lambda-GB-Second, Fargate, and BoxUsage (EC2 Instances).

  • Exclusion: NOT CONTAINS Spot.

---

Technical Observation and Governance

Unlike superficial analyses, this visualization refines eligibility through the usagetype field rather than just productname. This approach is technically superior because:

• Noise Filtering: It removes usage types that, although belonging to the service (such as data transfer fees or storage), are not eligible for the Compute Savings Plan.

• Spot Isolation: Ensures that Spot instances are removed from the calculation base.

• Focus on Committable Load: Ensures the final value presented is exactly the amount upon which the organization can apply coverage strategies.

---

Expected Insight

Daily evolution analysis should provide the answer for the investment "safety point":

• Identification of the Consumption Floor: Determine the minimum value the eligible cost reaches, even on low-utilization days (weekends or holidays). This "floor" is the ideal value for a 100% utilization commitment.

• Volatility Analysis: If there is significant daily variation (sharp peaks and valleys), the Savings Plan commitment should be more conservative to avoid coverage waste.

• Growth Trend: Identify if the baseline is rising or falling over the month, allowing for predictions on whether a new commitment will soon be necessary.

• Seasonality Validation: Differentiate structural consumption from sporadic consumption, ensuring the Savings Plan covers only what is permanent.

---

View - Service Cost by Period

Overview

This analysis aims to decompose eligible Compute Savings Plan consumption by AWS service. It identifies which workloads sustain the operational baseline and the predominant architectural profile of the environment (EC2 vs. Serverless), providing the necessary foundation for an accurate commitment strategy.

What the Analysis Considers

• Qualified Consumption: Only technically eligible recurring consumption.

• Data Purity: Strict exclusion of credits, adjustments, and non-recurring financial effects.

• Real Load: Considers exclusively the resources executed in the selected period.

• Proportional Distribution: Consolidation of cost by service for concentration and relative weight analysis.\

---

Selection Filters and Eligibility

• Line Item Type: lineitem/lineitemtype IS Usage

  • Restricts the analysis to effective resource consumption.

  • Excludes credits, refunds, support fees, financial adjustments, and lines related to Savings Plans or Reserved Instances already applied.

  • Ensures that the cost per service represents only real eligible consumption.

• Compute Services: product/productname IS (Amazon Elastic Compute Cloud, AWS Lambda, Amazon ECS or Amazon EKS).

• Refinement by Usage Type (lineitem/usagetype): Uses CONTAINS logic for the keys below AND NOT CONTAINS for Spot:

  • Lambda-GB-Second: Includes the effective execution of Lambda functions.

  • Fargate: Includes the consumption of serverless containers.

  • BoxUsage: Includes the use of On-Demand EC2 instances.

  • NOT CONTAINS Spot: Explicitly excludes Spot instances, which have dynamic pricing and do not compose the commitment baseline.

---

Technical Observation and Governance

The combined use of productname and usagetype is a high-governance practice in FinOps. This approach:

• Eliminates Distortions: Prevents ancillary fees (such as data transfer or disk storage) from inflating the commitment value.

• Ensures Eligibility: Isolates only the cost components that AWS allows to be covered with Compute Savings Plans.

• Focuses on Committable Cost: Ensures that the final value is the real amount subject to stable coverage.

---

Expected Insight

The analysis of the composition by service should guide strategic decision-making:

• Structural Service Identification: Determine which service sustains the core of the financial commitment.

• Architectural Predictability: EC2 instances tend to generate greater cost stability than highly elastic serverless workloads.

• Risk Concentration: Evaluate if high dependency on a single service might increase the risk of future variation in plan utilization.

• Commitment Calibration: Allow the FinOps team to define a purchase value proportional to the stability of the predominant workload (more aggressive in EC2, more conservative in Lambda/Fargate).

---

View - Average Cost per Hour

Overview

This analysis calculates the hourly baseline of consumption eligible for the Compute Savings Plan. It represents the primary indicator for sizing the financial commitment, translating consolidated consumption into a mean temporal metric that guides the purchase value per hour ($/hour).

What the Analysis Considers

To capture structural average behavior and avoid distortions, the analysis considers:

• Temporal Fidelity: Hourly granularity to capture intraday behavior.

• Qualified Consumption: Only recurring resources eligible for the plan.

• Noise Isolation: Exclusion of credits, adjustments, amortizations, and Spot consumption.

• Scope: The entire period selected in the workspace filters to ensure a statistically valid average.

---

Data Structure and Grouping

Dimension (Column)

• Usage Start Date (lineitem/usagestartdate): Represents the beginning of the consumption period recorded in the CUR.

  • Granularity: Grouped by HOUR. This grouping is essential for calculating the temporal average of eligible consumption.

Metric (Data)

• Unblended Cost (lineitem/unblendedcost): Gross value of consumption.

  • Aggregation (DATE_AVG): This function calculates the temporal average based on the date dimension, resulting in the average cost per hour throughout the entire analyzed period.

View Real Filters

• Line Item Type: lineitem/lineitemtype IS Usage.

  • Restricts the analysis to effective resource consumption.

  • Restricts the analysis to actual consumption. Excludes credits, support fees, and Savings Plans/RIs lines already applied.

  • Ensures that the hourly average is calculated only on real and committable consumption.

• Services: product/productname IS (Amazon Elastic Compute Cloud, AWS Lambda, Amazon ECS or Amazon EKS).

• Refinement by Usage Type (lineitem/usagetype): Uses CONTAINS logic for the keys below AND NOT CONTAINS for Spot:

  • Lambda-GB-Second: Lambda function execution.

  • Fargate: Serverless container consumption.

  • BoxUsage: On-Demand EC2 instances.

  • NOT CONTAINS Spot: Explicit exclusion of Spot instances so as not to inflate the commitment baseline.

---

Expected Insight

The hourly average is the "thermometer" for signing the Savings Plan contract:

• Baseline Determination: Defines the exact value of the safe range for hourly financial commitments.

• Stability Assessment: * If the hourly average shows low dispersion (stability): The commitment can be aggressive and approach the calculated average.

  • If there is high volatility (sharp peaks and valleys): The commitment should be conservative, focusing on the utilization "floor" to avoid coverage waste.

• Purchase Decision: The result allows for defining the value in dollars per hour that the organization is willing to commit with 100% utilization confidence.

---

View - Cost per Account

Overview

This analysis evaluates the distribution of eligible consumption among the organization's different AWS accounts. It is fundamental for measuring cost concentration and commitment governance risk, allowing an understanding of which accounts sustain the Savings Plan financial baseline.

What the Analysis Considers

To ensure a clear view of the organizational cost topology, the system considers:

• Consumption by Account: Recurring eligible value segmented by account ID.

• Noise Isolation: Exclusion of credits, adjustments, and non-operational financial effects.

• Baseline Purity: Explicit exclusion of Spot consumption, focusing only on committable loads.

• Full Period: Considers the complete time interval selected in the workspace to avoid seasonal distortions.

---

Data Structure and Grouping

Dimension (Column)

• Account ID (lineitem/usageaccountid): * Alias: Account Name.

  • Identifies the source AWS account of the consumption recorded in the CUR, allowing the cost to be mapped to the responsible business unit or environment.

Metric (Data)

• Unblended Cost (lineitem/unblendedcostt): Gross value of consumption.

  • Aggregation (SUM): Consolidated sum of eligible cost per account, with the total enabled for checking the organization's global volume.

View Real Filters

• Line Item Type: lineitem/lineitemtype IS Usage (Focus on real consumption).

  • Restricts the analysis exclusively to effective resource consumption per account.

  • Excludes credits, financial adjustments, administrative fees, and any Savings Plan or RI line already applied.

  • Ensures that the distribution by account represents only real eligible cost.

• Services: product/productname IS (Amazon Elastic Compute Cloud, AWS Lambda, Amazon ECS or Amazon EKS).

• Refinement by Usage Type (lineitem/usagetype): Uses CONTAINS logic for the keys below AND NOT CONTAINS for Spot:

  • BoxUsage: On-Demand EC2 instances.

  • Fargate: Serverless container consumption.

  • Lambda-GB-Second: Lambda function execution.

  • NOT CONTAINS Spot: Explicit exclusion of Spot instances.

---

Expected Insight

1. Map the origin of eligible consumption:

   • Sustains the definition of the global Savings Plan commitment.

2. Evaluate recurring baseline concentration:

   • Relevant architectural changes in accounts with high volume can directly impact the commitment utilization rate.

3. Understand Savings Plan absorption predictability:

   • The more concentrated the eligible consumption, the greater the coverage predictability.

   • The more distributed, the greater the dependence on multiple accounts to maintain high utilization.

4. Support financial Chargeback strategy:

   • As AWS automatically applies the benefit where there is the greatest eligible discount, the actual distribution of savings is not defined by the customer.

   • This view guides the definition of internal rules for apportionment and redistribution of the benefit among areas.

5. Dispersion Analysis:

   • High concentration: Lower operational complexity and higher coverage stability.

   • High dispersion: Requires active governance to monitor global utilization and define a clear benefit-sharing policy.

6. Assist in deciding the commitment level:

   • Environments with a fragmented but stable baseline tend to support more robust commitments when analyzed on a consolidated basis.

---

View - Estimated Savings – No Upfront – Compute Savings Plan

Overview

This analysis performs a simulation of the financial impact of adopting the Compute Savings Plan in 1-year and 3-year scenarios (No Upfront model). It presents potential savings segmented by workload, serving as the most granular and detailed view available to justify investments and financial commitments.

What the Analysis Considers

To ensure the financial projection is realistic and conservative, the system applies:

• Qualified Consumption: Consolidation of recurring eligible consumption.

• Multidimensional Segmentation: Analysis by account, service, region, operating system, and instance type.

• Volatility Exclusion: Explicit removal of Spot consumption from the calculation.

• Real Discount Factors: Application of estimated average market rates (27% for 1 year and 48% for 3 years).

• Data Consistency: Full coverage of the period selected in the workspace.

---

Table Structure (Dimensional Columns)

This visualization uses a table structure to allow cross-correlation between the following dimensions:

Column	CUR Attribute	Description
AWS Account	usageaccountid	Identifies the organizational unit (Account Name).
Service	productname	Indicates the specific workload (EC2, Lambda, ECS, EKS).
Region	region	AWS Location (important for capturing regional price variations).
O.S.	operatingsystem	Operating System segmentation (e.g., Linux vs Windows).
Instance Type	instancetype	Identifies specific hardware (e.g., m5.large, t3.medium).

---

Simulation Metrics and Formulas

The metrics calculate the financial benefit by comparing the current scenario with savings projections:

• Current Cost (unblendedcost): Represents the amount spent today in the eligible On-Demand model.

• Savings Plan – 1 Year: Calculated via custom formula:

  $$Current \ Cost \times 0.27$$

  Simulates an estimated average savings of 27%.

• Savings Plan – 3 Years: Calculated via custom formula:

  $$Current \ Cost \times 0.48$$

  Simulates an estimated average savings of 48%.

---

View Real Filters

1. Line Item Type: lineitem/lineitemtype IS Usage

   • Limits the financial simulation to effective On-Demand consumption.

   • Excludes credits, discounts already applied, amortizations, and financial adjustments.

   • Avoids double counting of benefits and ensures that the savings projection is calculated on the correct base.

2. Services: product/productname IS (EC2, Lambda, ECS, EKS).

3. Usage Type Refinement:

   • Logic: CONTAINS (Lambda-GB-Second, Fargate, BoxUsage) AND NOT CONTAINS (Spot).

   • Objective: Ensures that only structural On-Demand consumption is considered in the simulation.

---

Expected Insight

This view is the fundamental piece for the Savings Plan "Business Case":

• ROI Quantification: Translates the technical baseline into monetary values of potential savings.

• Sensitivity Analysis: Compares the financial gain between short-term (1 year) and long-term (3 years) commitments.

• Leverage Point: Identifies which Instance/Region/O.S. combinations offer the highest return potential.

• Executive Decision Support: Provides a structured and granular financial argument for budget approvals and spending commitments.