Optimization

Optimization recommendations records a resource identifier, proposed action, estimated monthly savings, review status and verified savings. It gives the service owner a place to discuss the proposed action before committing to a change. The recommendation can refer to an operational incident so the team can see why it was raised.

A program can group repeatable optimization work for a service, while a rollout records its scope, coverage, approval and activation. A verification record compares target and realized savings for a workload in a cluster. These are different decisions: whether an idea looks worthwhile, whether a rollout may proceed, and whether the observed result met the target.

• Candidate. Record the affected resource and a specific saving action.
• Decision. Move the recommendation through review and approval before applying it.
• Evidence. Compare realized savings with the target after the workload has run.

Choose a baseline period for each saving estimate. Name who may review and approve a recommendation, and whether that person can also approve its related change. For a program rollout, define the scope in terms the service owner can inspect and record the person who approved it. Decide when a verification is due, because a change made yesterday may not yet have a representative cost period.

Where a purchased commitment is involved, link the entitlement to the asset using it and keep the assignment and release dates. This helps the team avoid claiming a saving twice when commitment coverage changes.

A utilization sample shows an underused asset. The Service Owner records a recommendation to reduce its capacity and an estimated monthly saving. Finance Manager checks the estimate against daily charges. After review and approval, the owner prepares a service change with a rollback plan. The Cloud Operations Engineer records the deployment outcome for the service. The owner compares the affected workload and time period with the recommendation, then records target and realized savings and updates verified savings on the recommendation. If realized savings are lower than expected, the team keeps that difference visible and checks for a reversal at the next checkpoint.

• Recommendation Board — Follow ideas from proposed through applied, with estimates and results.
• Change Schedule — See when approved work is planned.

Continue with change to verification and spend review.