INFRASTRUCTURE LEARNINGINFRASTRUCTURE LEARNING
インフラ大全インフラ大全IT INFRASTRUCTURE / LEARNING
Infrastructure learning guide
Find a lesson or lab/

Esc to close · Search terms stay on your device

LEARNING日本語

Topic lessonTopic: Bill increases and data transfer costs

Unexpected AWS bill increase: investigate services, usage types and retained resources

Start an AWS bill investigation by aligning the period and cost basis, then locate the line items that increased. Separate service, Region, usage type, quantity, rate and adjustments. A fictional cost table also helps explain why charges can remain after a server is stopped.

What you will learn

Align comparisons, identify increased cost components and quantities, and choose checks that account for retained resources and reporting delays.

Prerequisite

Billing views support invoice/payment review; Cost Explorer supports analysis. Current-month estimates differ from finalized bills. Amounts here are hypothetical arithmetic, not a live quote or an operation that stops charges.

Align the period and meaning of the costs

Settings to align before interpreting a difference
ConditionRisk of a mismatch
Account/organization and scopeMistaking another account’s usage for one resource’s growth
Dates, duration and estimate/final statusComparing a partial month with a whole month or different durations
Cost basisConfusing discount/reservation allocation, credits, refunds or taxes with usage growth
Data update timeTreating reporting delay as proof that unchanged usage continues after stopping something

Cost Explorer refreshes cost data on a daily basis, and upstream data can arrive later than 24 hours. Billing views can differ in update timing and grouping. An immediate post-stop view proves neither a failed stop nor zero remaining charges.

Narrow from increased line items to subjects and usage types

A cost investigation order, not a resource-deletion procedure
A cost investigation order, not a resource-deletion procedure1: Comparison basis — Compare differences on the same basis — Increased components. 2: Increased components — Inspect the subjects behind the increase — Quantity and subject12Comparison basisIncreasedcomponentsQuantity andsubjectA cost investigation order, not a resource-deletion procedure1: Comparison basis — Compare differences on the same basis — Increased components. 2: Increased components — Inspect the subjects behind the increase — Quantity and subject12ComparisonbasisIncreasedcomponentsQuantity andsubject
  1. Comparison basis → Increased componentsCompare differences on the same basis
  2. Increased components → Quantity and subjectInspect the subjects behind the increase
Comparison basis
Aligned period and cost basis
Increased components
Service, Region and usage type
Quantity and subject
Usage, rates and retained resources
  1. Confirm the period and estimate/invoice status in the relevant billing view.
  2. Identify increased services and accounts in cost analysis.
  3. Inspect Region, usage type and available operation/resource breakdowns.
  4. Separate quantities, rates and adjustments while checking units.
  5. Review ongoing use, retained resources and plan/commitment terms before changing anything.

B1: calculate changes in quantity and cost

B1 assumptions: do not sum quantities with different units
ComponentUnitPreviousCurrentInvented rate
ComputeHours100200$0.05/hour
StorageGB-month5050$0.10/GB-month
TransferGB10100$0.20/GB
B1 calculation: components of the $23 increase
ComponentPrevious costCurrent costDifference
Compute$5$10+$5
Storage$5$5$0
Transfer$2$20+$18
Total$12$35+$23

Transfer contributes the largest B1 increase, $18: 90 additional GB at the invented $0.20 rate. Together with $5 for compute, the total increase is $23. This table alone does not identify whether increased transfers came from demand, caching, path changes, retries or another cause.

Distinguish stopping a server from ending related charges

Related subjects to review after stopping or terminating
SubjectExample reason charges remainReview
EBS volumes/snapshotsStorage and configuration can outlive computeReview need, retention, dependencies and billed components
NAT and other networkingBilled time or processing can continueReview mode, AZ-hours, actual paths and related costs
S3 versions/incomplete uploadsStored data can remain beyond current visible objectsInspect versions, incomplete work and retention conditions
Other Regions or service-managed resourcesUsage can continue independently of the stopped subjectVerify scope and the managing service before changes

Disabling a Region is not the same as deleting its resources or ending all charges. Review operational need, retention obligations, dependencies and authority before removing apparently unused resources. “Delete everything unknown” can destroy investigation and recovery evidence.

Review whether resources are still needed, even when costs have not risen. Use the component table above to check each resource's owner, purpose and retention requirements. The fictional cases below have unchanged costs on the same comparison basis.

Fixed case: unchanged costs, different reasons to retain or review
Component and resourceVerified facts in this exampleNext decision
Compute: standby EC2 instanceThe owner approves its standby role during failures. Normal traffic is low.Retain it to meet that requirement and set the next review date.
Storage: recovery snapshotThe retention policy in this example still requires keeping it.Record the reason and deadline; reassess recovery needs after that deadline.
Storage: EBS from a completed testThe owner confirms the test has ended. No required data, retention obligation or dependency remains.Treat it as a change candidate; check authority and impact before changing it.

Choose an option to see the result and explanation.

Change the conditions: the test EBS volume's purpose and retention requirements are unknown, and its cost is unchanged. What next?
Read the answer and reason

Answer: Pause changes until purpose and retention are confirmed · Those facts were verified in the table; here they are unknown. Confirm them with the owner before reviewing authority and impact.

Record why resources are retained and when to review them again, or what changed and the outcome. After cost data updates, compare the affected components on the same basis and check that operational and recovery requirements are still met.

Choose the next check from a cost increase

Choose an option to see the result and explanation.

Which components account for B1’s $23 increase?
Read the answer and reason

Answer: Compute +$5 and transfer +$18 · The differences are 10−5 and 20−2, totaling $23; compute each quantity in its own unit.

Can you directly add usage quantities measured in hours and GB?
Read the answer and reason

Answer: Analyze quantities and costs by unit · Compare each usage type using its unit and billing conditions.

An EC2 instance is stopped. Must every related charge be zero?
Read the answer and reason

Answer: No; review storage, networking and other components · Stopped compute and retained EBS, NAT or other subjects can have separate charges.

Costs appear higher immediately after stopping something. What should you check first?
Read the answer and reason

Answer: Update time, period and remaining components · Reporting can lag and related resource charges can continue separately.

Check, then explain in your own words

  1. Verify B1’s $12→$35 change and $23 difference.
  2. Separate usage units from summed monetary costs.
  3. Review retained components and data freshness after stopping resources.
Check your understanding: What is the first sequence for a bill increase?

Align the comparison, identify increased components, then separate quantities, rates and adjustments.

This separates reporting/comparison differences before investigating resources and paths.

Sources and conditions checked

The official references used in this lesson are collected here with their scope, conditions and corresponding sections. Independent expert review and live AWS comparison have not been performed.