Cost insight financial metrics
DoiT cost insights can display up to four financial metrics on each recommendation. This page explains what each metric means, how the figure is calculated, and how to read them together when you prioritize what to fix first.
The metrics cover both your current spend and your commitments (Savings Plans, Reserved Instances, and Committed Use Discounts), for example, whether acting reduces your bill, frees committed coverage, or prevents a future charge.
| Metric | Answers the question | Direction |
|---|---|---|
| Potential cost savings | How much will I stop paying if I act? | Reduces current spend |
| Cost avoidance | What future charge do I prevent by acting in time? | Prevents an upcoming increase |
| Freed commitments | How much committed spend gets freed up to reuse? | Frees existing commitment |
| Avoidable commitments | What extra commitment or on-demand cost do I steer clear of? | Prevents future commitment waste |
Data sources
The financial metrics are calculated from several sources so the figures reflect what you actually pay, not list price:
-
Your billing data: Establishes the current baseline cost of the resource and your effective discount rate.
-
Your contracts and price books: Applies your negotiated discounts to the target (after-action) price.
-
Cloud provider pricing catalogs: Supplies current list prices for the target configuration.
-
Commitment-aware logic: Detects whether a resource is covered by a Savings Plan, Reserved Instance, or Committed Use Discount (CUD), and which type, so the effect on your commitments is calculated separately from the effect on usage.
Because usage cost and commitment cost move independently, a single action can affect more than one metric at once. See Combined metrics.
Values are estimates based on recent billing data and current pricing catalogs. Actual results depend on when and how you act and on later usage changes. Numbers are rounded for display; for more detail, use the insight details page and the underlying reporting.
Potential cost savings
Potential cost savings is the estimated amount you would stop paying by acting on the insight — money that comes off your current bill. On the Insights table this appears as Cost savings; on overview cards and insight metadata it often appears as Potential cost savings or Potential savings.
The metric compares the baseline cost of the affected resource over a recent period (adjusted for your discounts) to the cost of the recommended end state. For a deletion, the end-state cost is zero, so the saving equals the resource's current cost. For a downgrade or migration, it's the difference between the current and the smaller or cheaper target configuration.
For example, an idle load balancer costs $42 per month. The recommendation is to delete it. Potential cost savings = $42 per month — that amount leaves your bill once you delete it.
Today this metric reflects savings from removing idle or unused resources. Rightsizing and service-migration savings (for example, gp2 to gp3, or moving to a more cost-efficient instance family) are being added and will also populate this field.
Cost avoidance
Cost avoidance is a future charge you can prevent by acting before it starts — most commonly end-of-life (EOL) or extended-support fees. Unlike Potential cost savings, this isn't money you're paying today; it's money you would start paying if you did nothing. On some surfaces this appears as Potential cost avoidance.
The extended-support or penalty rate is taken from the pricing catalog (with your discounts applied), along with the date it takes effect, to determine the charge you avoid by upgrading before that date.
For example, a managed database engine version reaches end of standard support in two months, after which extended-support fees of $180 per month begin. Upgrading before the EOL date avoids that charge. Cost avoidance = $180 per month starting at the effective date.
Freed commitments
Freed commitments reflect existing commitment spend (Savings Plan, Reserved Instance, or CUD) that currently covers the affected resources and would be freed up if you act. That freed coverage can be applied to other eligible usage now, or it lets you commit to a lower amount at your next renewal.
Acting on the insight does not cancel a commitment you've signed. It releases the coverage so it can be used more effectively elsewhere.
The portion of your commitment currently absorbed by the resource is identified and reported as the amount freed by the recommended action.
For example, an idle instance is covered by a Savings Plan worth $87 per month. Deleting the instance saves the on-demand or usage portion (Potential cost savings) and frees the $87 per month of Savings Plan coverage it was consuming. Freed commitments = $87 per month.
For example, you downsize an instance from a larger to a smaller type, both covered by a Savings Plan. Because you've already committed to the spend, there is no immediate cash saving on the compute (Potential cost savings = $0), but the commitment coverage required drops: current $87 per month − target $22 per month = Freed commitments = $65 per month.
How much immediate value you get from freeing coverage depends on the commitment type:
| Commitment type | Behaves like | When you release a resource it covers |
|---|---|---|
| Compute Savings Plan, flexible or spend-based CUD | A dollars-per-hour promise, not tied to one resource | Coverage re-applies automatically to other eligible usage. Real value now. |
| Standard Reserved Instance, resource-based CUD | Tied to a specific instance family and region | Coverage does not move; it keeps billing until expiry. Free-up value mainly at expiry — acting early can create an avoidable cost. |
Freed commitment becomes a realized saving only when the freed coverage is either re-applied to other eligible usage or reduced at renewal. Flexible commitments (Compute Savings Plans, flexible CUDs) re-apply to other eligible usage automatically. Inflexible commitments (Standard Reserved Instances, resource-based CUDs) free up only when they expire.
Avoidable commitments
Avoidable commitments represent additional commitment consumption — or on-demand and unused-commitment cost — that you would otherwise take on, and can avoid by acting on the insight before your next commitment decision.
The commitment or penalty impact of the current environment is modeled to report the extra commitment spend (or on-demand cost from an inflexible commitment) you steer clear of by taking the recommended action or timing it correctly.
For example, a resource is covered by an inflexible Standard Reserved Instance worth $120 per month. Downsizing it now would not save money — the Reserved Instance keeps billing in full until it expires, and the new smaller resource runs on demand, so you would pay for both. By timing the change to the Reserved Instance's expiry, you avoid that overlapping cost. Avoidable commitments = $120 per month until expiry.
$0 is a common and correct value for this metric. When the recommended actions are all "delete idle" or "free existing commitment," there is no additional future commitment to avoid, so Avoidable commitments is $0. This does not indicate missing data. A dash (—) means the value could not be retrieved.
Combined metrics
A single action often affects more than one metric, because usage cost and commitment cost are separate. Reading them together tells the full story:
-
Idle resource, not covered by a commitment → Potential cost savings only.
-
Idle resource, covered by a Savings Plan → Potential cost savings (usage) and Freed commitments (freed coverage).
-
Rightsizing under a Savings Plan → little or no Potential cost savings, but a Freed commitment (lower coverage needed).
-
End-of-life upgrade → Cost avoidance (prevented future fee).
-
Change against an inflexible Reserved Instance → may show an Avoidable commitment rather than a saving, if acting early would cost more.
Don't add Potential cost savings and Freed commitments together as if they were the same money — they measure two different things (bill reduction vs freed commitment coverage). Treat them as complementary signals when you prioritize.
Metrics in the console
All amounts are displayed in your organization's currency. Each metric has a hover tooltip on the Insights dashboard and the insight details page that shows how the number was calculated for that specific insight.
On the insight details page, these four metrics appear in the metadata section alongside fields such as Source and Severity. On affected resources, the same metrics appear as table columns.
The columns, fields, and time frames display dynamically based on your current view. The Insights table uses monthly amounts and short column names such as Cost savings and Cost avoidance, while overview cards and insight metadata often show annualized figures and labels such as Potential cost savings, Potential savings, Potential cost avoidance, Freed commitments, and Avoidable commitments. A dash (—) indicates that data is not available for that insight.
Not every metric applies to every insight. A metric that isn't relevant to a given recommendation may show as $0 or as a dash (—). For example, an insight about deleting an idle resource typically shows Potential cost savings but not Cost avoidance.
You can view financial metrics in the following places:
-
Insights dashboard: Overview cards can show metrics such as Potential cost savings and Freed commitments as annualized amounts when you view the Cost or All categories.
-
Insights view: Each metric appears as a sortable column (Cost savings, Cost avoidance, Freed commitments, Avoidable commitments), so you can rank insights by the impact that matters most to you.
-
Insight details: The metrics appear in insight metadata and on each affected resource, with hover tooltips that explain the calculation for that insight.
-
Threads: When you create a thread from an insight that has Potential cost savings, the estimated monthly savings are included automatically.