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: AWS billing analysis and cost optimization

Verify Compute Optimizer recommendations against SQL Server peak workloads

A lightly loaded host on an ordinary day may need memory and I/O at month end. Tie a size recommendation to its observation window and a representative SQL Server task.

What you will learn

Tie a size recommendation to its observation window and a representative SQL Server task.

Prerequisite

This exercise uses fictional evidence and makes no changes in a live environment.

From recommendation to acceptance

Inspect Compute Optimizer recommendations, performance risk and utilization history. Exercise monthly close and memory-sensitive work.

A repeatable resize decision
A repeatable resize decision1: Recommendation — Inspect observations — Workload. 2: Workload — Repeat the critical workload — Test. 3: Test — Compare acceptance limits — Decision123RecommendationWorkloadTestDecisionA repeatable resize decision1: Recommendation — Inspect observations — Workload. 2: Workload — Repeat the critical workload — Test. 3: Test — Compare acceptance limits — Decision123RecommendationWorkloadTestDecision
  1. Recommendation → WorkloadInspect observations
  2. Workload → TestRepeat the critical workload
  3. Test → DecisionCompare acceptance limits
Recommendation
Candidate and lookback
Workload
Close process and memory
Test
Use equivalent synthetic data
Decision
Check time and cost

Same synthetic month-end workload

The allowed completion time is 15 minutes. These are fictional observations.
MeasurementCurrentSmaller candidate
Normal CPU20%35%
Close runtime12 minutes20 minutes
I/O waitsLowHigher
DecisionWithin limitHold even if cost would fall

If the observation window misses month-end, exercise it separately before accepting the recommendation.

Verify recommendation assumptions

Read the history Check the recommendation, risk, observation period and available metrics.

Hold workload conditions fixed Compare runtime, I/O, memory, representative dates and licensing terms.

Verify the change Record acceptance, rollback conditions and later billing reconciliation.

Check your decision

Choose an option to see the result and explanation.

The close takes 20 minutes and the limit is 15. What now?
Read the answer and reason

Answer: Hold and inspect I/O and memory · A representative task exceeds the limit.

The observation window omits month-end. What do you do?
Read the answer and reason

Answer: Exercise monthly close separately · Add a representative workload measurement.

What belongs in the cost comparison?
Read the answer and reason

Answer: Runtime, licensing and eligible discounts · Use the applicable charging terms.

Check, then explain in your own words

  1. Inspect the observation window.
  2. Measure representative close behavior.
  3. Accept only when operating limits pass.
Check your understanding: Can you implement a recommendation as soon as it appears?

Test critical tasks under equivalent conditions and meet the operating limits.

A recommendation and a workload requirement measure different things.

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.